Sie sind auf Seite 1von 79

Session 1: Long-span structures for building and spaces: materials and concepts

Objekttyp:

Group

Zeitschrift:

IABSE reports = Rapports AIPC = IVBH Berichte

Band (Jahr): 71 (1994)

PDF erstellt am:

05.08.2018

Nutzungsbedingungen Die ETH-Bibliothek ist Anbieterin der digitalisierten Zeitschriften. Sie besitzt keine Urheberrechte an den Inhalten der Zeitschriften. Die Rechte liegen in der Regel bei den Herausgebern. Die auf der Plattform e-periodica veröffentlichten Dokumente stehen für nicht-kommerzielle Zwecke in Lehre und Forschung sowie für die private Nutzung frei zur Verfügung. Einzelne Dateien oder Ausdrucke aus diesem Angebot können zusammen mit diesen Nutzungsbedingungen und den korrekten Herkunftsbezeichnungen weitergegeben werden. Das Veröffentlichen von Bildern in Print- und Online-Publikationen ist nur mit vorheriger Genehmigung der Rechteinhaber erlaubt. Die systematische Speicherung von Teilen des elektronischen Angebots auf anderen Servern bedarf ebenfalls des schriftlichen Einverständnisses der Rechteinhaber.

Haftungsausschluss Alle Angaben erfolgen ohne Gewähr für Vollständigkeit oder Richtigkeit. Es wird keine Haftung übernommen für Schäden durch die Verwendung von Informationen aus diesem Online-Angebot oder durch das Fehlen von Informationen. Dies gilt auch für Inhalte Dritter, die über dieses Angebot zugänglich sind.

Ein Dienst der ETH-Bibliothek ETH Zürich, Rämistrasse 101, 8092 Zürich, Schweiz, www.library.ethz.ch

http://www.e-periodica.ch

#4 n SESSION 1 LONG-SPAN STRUCTURES FOR BUILDING AND SPACES: MATERIALS AND CONCEPTS
#4
n
SESSION 1
LONG-SPAN STRUCTURES FOR BUILDING AND SPACES:
MATERIALS AND CONCEPTS
Lee Bla

Lee

Bla

Jörg SCHLAICH Prof. Dr Univ. of Stuttgart Stuttgart, Germany Jörg Schlaich, born 1934, studied in
Jörg
SCHLAICH
Prof.
Dr
Univ. of Stuttgart
Stuttgart, Germany
Jörg Schlaich, born 1934,
studied in Stuttgart, Berlin
and Cleveland, USA. De¬
signer of
bridges,
buil¬
dings,
long
span
light-
weight structures, cables
and solar
energy, resear-
cher in structural
concrete.
Director of
Institute for
Structural Design, Univ. of
Stuttgart, and
Partner of
Schlaich, Bergermann and

Conceptual Design o

Concept et projet de toit

Konstruktiver Entwurf w

Partner, Consulting Eng.

14 CONCEPTUAL DESIGN OF LONG-SPAN ROOFS 1. DESIGN PRINCIPLES In designing long-span structures engineers first
14 CONCEPTUAL DESIGN OF LONG-SPAN ROOFS
1. DESIGN PRINCIPLES
In designing long-span structures engineers first of all must get rid of their habit to think in fix
proportions, forgetting about the dominating
influence of scale. It is misleading to give the re-
quired thickness d of a beam or girder as a
fixed fraction d/L of its span L with d/L =1/18
for a Single span and 1/22 for multiple span or so. No: Simply by designing a beam or slab
with rectangular cross-section to carry its dead load only we find:
y: density of the material
ß: strength of top and bottom
F 2
# ?*
fibre in tension and compression
Ljjm: limit span of a beam under its own weight
d/Lreq: required depth/L|jm
M b-d
ybd-L2
->
>-±Lm*const.
^
(see fig.
1)
8 b
d:
4
ß
d/L is not a constant figure but increases linearly with the span and depends further on the
efficiency of the material y/ß; the lighter the material and the higher its strength, the smaller
will be the required thickness d of a beam of a given span. But its proportions will change
dramatically with its absolute size, a fact which can be well observed frorn natural structures
(fig. 1).
d/Lreq A
ooa
[km]
Fig. 1 The role of scale in natural structures [1][2]
With increasing span, solid girders become more and more clumsy, their dead load eats away
their strength and for long-span structures something has to be done to by-pass this awkward
Situation.
The answer is of course well known: We have to leave
away
all that material of our beam
which is not fully used: at mid span, we need to keep only the top and bottom fiber and leave
away the web and towards the Supports we can reduce the chords but need the web. Thus,
via T-, TT-, U-girders and hollow slab- and box-girders, we reach the different types of
trusses (fig. 2, top):
Their main feature is that they avoid bending and carry loads only by axial forces, thus mak-
ing füll use of the material's strength and reducing their dead load to a minimum.
J. SCHLAICH, R. BERGERMANN 15 Bending The next Step towards lightness for increasing spans is
J. SCHLAICH, R. BERGERMANN
15
Bending
The next Step towards lightness for increasing
spans is to subdivide the girder into a primary
structure acting either in compression (fig. 2,
bottom left) or in tension (fig. 2, bottom right).
aaaaa
The primary structure in compression (tension)
may either attribute its tension- (compression-)
AAAAAAAA&
partner to the subsoil - in that case we speak of
an earth-anchored or a true arch (Suspension)
bridge - or utilize its secondary structure (usually)
for self-anchorage.
Tension
Cotuprcssio
An earth-anchored structure is (usually) easier to
erect but costlier, whereas for a self-anchored
.^ ^- *^ -^
just the opposite is true. The cable-stayed
system combines both advantages: it is self-
anchored (resulting in cheaper foundations) and
can be constructed free cantilevering without
temporary falsework (fig. 3). Beyond its self-
balancing function the purpose of the secondary
system is to stiffen the primary system and to
serve as an envelope in case of a building or as a
road/railway in case of a bridge.
In case of a building - the theme of this Sym¬
posium - the envelope needs not necessarily be
straight or horizontal, as it does in case of a
bridge, but also there these hybrid structures
(M. Saitoh speaks of Beam-String-Structures BSS
in case of the primary structure acting in tension)
are functionally better adapted to the required
space as double-curved space structures, which
we shall discuss further below (fig. 4)[3J.
Fig. 2 The development of the girder
The primary structure, especially if it acts in
compression, must be stabilized or stiffened by
the secondary structure. Making use of all
possibilities to stiffen the primary system
- the weight of the structure as a whole
- the bending stiffness of the girder (and the
arch)
- the geometry of the primary system (triangulär
or quadrangular mesh)
- overall prestress (only applicable to cable
bridges)
- and even combining them in an intelligent way
T
-a_-.
_^i_
- not speaking of combining different materials in
the same bridge
Fig. 3 The earth-anchored suspended girder needs abutments but can be built without false¬
work. The self-anchored girder needs temporary falsework but avoids horizontally loaded
foundations. The cable-stayed girder combines both advantages.
16 CONCEPTUAL DESIGN OF LONG-SPAN ROOFS Stands for the intellectual appeal and the joy of
16
CONCEPTUAL DESIGN OF LONG-SPAN ROOFS
Stands for the intellectual appeal and the joy of
structural design. Some examples of the be-
haviour of hybrid structures are given in figure 5.
^
The further we proceed down along the list of
girder types, as sketched in figure 2, the more
flexible they become, but we simultaneously
V
realize that the types further up in the list have to
pay for their higher stiffness, resulting frorn
triangulär mesh, with improved fatigue strength
and ductility of their structural elements,
especially their cables.
Fig. 4
Hybrid structures adapt better to the required space than double-curved structures
which usually encase too much volume.
"TrrTTTr^-a^^rTrrrTr
i jTrrrr^.^Jiu^rrrrTTf
"iimniiUH
niilliitiitiiliiiiiiiiMiiiMHii.il
»luunHiimi
Mi iiiiuni im mti
— »
H"
~-
2J
ooa»
4
10J
oooo»
Fig. 5 Comparison of deformations and frequencies of suspended girders with vertical hangers
(H"), inclined hangers (Hx) and of cable-stayed girder (S) with varying moment of inertia J of
the girder.
^
Ma
' IIIIIIIIIIJ
g ~ T
„|
i
k
^-u_i
rx^
Fig. 6 Hybrid tension structures for an ice-skating
hall at Memmingen (left, several alternatives) and
for a multi-purpose hall at Karlsruhe [3]
J. SCHLAICH, R. BERGERMANN 17 If we are not too timid as far as deformations
J. SCHLAICH, R. BERGERMANN
17
If we are not too timid as far as deformations are concerned, we can make use of this whole
catalogue also for long-span building roofs, as many examples recently built demonstrate
(fig. 6). We are inclined to call this type of buildings "High Tech Architecture", insinuating
that they are an invention of our times. This is not so as a Stadium roof proposed as early as
1927 by Heinz and Bodo Rasch frorn Stuttgart clearly demonstrates (fig. 7).
rrt^r
y /
m
Fig. 7 Stadium project by Heinz and Bodo Rasch
With these hybrid structures the load bearing and the enveloping functions are usually inde-
pendent or additive. As mentioned, this may be an advantage as far as the encased volume is
concerned (fig. 4) but of course if both functions are combined, the overall result must be
more efficient. This brings us to the double-curved surface structures (fig. 8).
Similar to the girders (fig. 2), they either combine compression and tension or they work pri-
marily in compression or in tension. And again, proceeding down the list, they become in-
creasingly flexible or deformable, depending on the type of stiffening respectively the topolo-
gy of the surface
- the continuous surface
- with triangulär mesh
- with quadrangular mesh
At their base, both lists (fig. 2 and fig. 8) meet with the same unstiffened arch or catenary
cable.
Of course, beyond these
pure
surface structures
Li j i ii\\ TTD
there is again the possibility of combining them
1 ~7
4
with other structural members, e.g. the Shells
may be stiffened by cables in the shape
of
BENOjING
spokes wheels or the cable nets by interaction
/WW\
with a roofing which avails of some shell action.
-COMPRESSION. TENSION -
We further recognize again prestress,
mechanically applied to surfaces with anticlastic
curvature or pneumatically applied to surfaces
with synclastic curvature.
And again, there is a dose interrelation between
the topology of these structures and their con¬
struction method, usually prefabrication and
3
LAYER
CABLE
NET
erection: The concrete Shells, though most effi¬
cient structures, and the very concrete struc¬
tures, suffer frorn their costly frameworks, since
double-curved surfaces are non-developable.
MD 2
LAVER
CABLE
NET
Several efforts have been made to revive con¬
crete Shells. Heinz Isler is most successful with
that, as demonstrated by his beautiful Shells.
CATENARY
CA8LE
Fig. 8 The development of
Another approach is the use of pneumatic form-
work (fig. 9). More common than concrete
double-curved surface structures
Shells are today the spherical grid domes in their
18 CONCEPTUAL DESIGN OF LONG-SPAN ROOFS different topologies, amongst them best known the geodesic dorne
18 CONCEPTUAL DESIGN OF LONG-SPAN ROOFS
different topologies, amongst them best known the geodesic dorne [5]. The problem is how
to cover a double-curved surface with triangulär mesh with as many as possible elements,
struts and nodes, of equal geometry (see example 1 in section 2 of this paper).
On the tension side of figure 8 we face in principle the same manufactural problem. Some
basic answers are compiled in figure 10.
STRUCTURE
MANUFACTURE
GEOMETRY
free
SQUARE NET
restncted
TRIANGULÄR
NET
free
Vät
TEXTILE
MEMBRANE
5=SS
restncted
JFx
THIN METAL
SHEET
MEMBRANE
Fig. 9 Pneumatic formwork for concrete
Shells: suitable shapes [4]
Fig. 10 Double-curved tension structure:
The interrelation between type of structure/
manufacture/geometry or load bearing behaviour
The 2-layer cable net with quadrangular mesh is easy to manufacture and permits a variety of
forms, unmatched by any
other structural type, but it has to pay with large forces, large de-
formations and finally
high costs. The Munich Olympic Roof (1972) is up to date the largest
application [6]. The 3-layer cable net has an ideal membrane-shell load bearing behaviour but
out of manufactural reasons, its geometry is limited to rotational shapes. The cable-net
cooling tower at Schmehausen (1975) was the largest application [7]. Textile membrane
structures can be completely prefabricated in the shop using a cutting pattern, are brought to
the site, unfolded there and erected without any falsework. They can be combined with pri¬
mary cable structures (as shown by example 2 of this paper). Convertible roofs are the high
art of membrane structures [8]. Finally metal membrane structures, rarely applied up to now,
can either be made frorn strips and welded on site, as done for the Moscow Olympic Struc¬
tures [9], or thin stainless steel sheets can be pneumatically deformed utilizing their immense
plasticity, as demonstrated with the manufacture of solar concentrators [10].
The authors are aware that the above was only a short summary
of basic design principles of
long-span roofs. But instead of filling all pages allotted for this
paper with abstract considera-
tions, a short presentation of two structures, recently designed by the authors, may be more
illustrative.
J. SCHLAICH, R. BERGERMANN 19 2. TWO EXAMPLES: A GLASS COVERED GRID SHELL AND A
J. SCHLAICH, R. BERGERMANN
19
2. TWO EXAMPLES: A GLASS COVERED GRID SHELL AND A
HYBRID MEMBRANE
STRUCTURE
2.1 The glass roof over a courtvard of a museum in Hamburg
Glass roofs are attractive frorn an architectural as well as climatical point of view. Having al-
ready been the symbol of the new architecture of the Industrial Revolution during the 18th
and 19th Century, they experienced a revival during the second half of this Century through
the work of pioneers like Walther Bauersfeld, Konrad Wachsmann, Buckminster Füller, Max
Mengeringhausen, Frei Otto and others.
Obviously the most favourable basis for a translucent roof is the double-curved reticulated
spatial structure with triangulär mesh. Such structure, however, especially if directly glazed
without intermediate glass frames, evokes three basic problems:
- The geometrical problem to cover a double-curved, i.e. non-developable surface with tri-
angles, having for manufactural simplicity as many as possible members and nodes of
equal size. (This problem obviously got some relief through recent progress in CNC-
manufacturing.)
- Glass panels are preferably produced in quadrangles, of course permitting a Variation of
their angles. Therefore only two out of three members of the triangle constituting the
structure should support the glass.
- Especially for double glazing the quadrangular glass panels must either be produced double-
curved to fit the structure's surface, or the geometry of the structure must be chosen so
that the four node points of each mesh are in one plane and may be glazed with plane
glass panels. For Single glazing, however, some warp of the glass panels is acceptable.
It is impossible to discuss this whole issue and all possible Solutions here. Therefore, one So¬
lution, recently developed by the authors, shall be presented and exemplified [11]. The aim is
to arrive at a structure corresponding to what is called "space dorne" in figure 8 but suitable
to be built for non-mathematical free shapes, too.
The basic grid of the structure when developed into a plane is a square net consisting of flat
bars (fig. 11). This plane Square net may be turned into almost
any
type of shape by changing
the original 90° mesh angle. The quadrangles become rhomboids. This way any double-
curved structure suddenly becomes "developable", and accordingly simple is the assembly of
the basic grid. It entirely consists of bars, identical in length, bolted together, pivoting at their
intersections. Bars of different length occur only at the outer edge as dictated by the structu¬
ral geometry. The mesh angles are determined by the intended structural shape.
#
^
R
9
2-K
-R
<p-
a) View
b) Elevation (with diagonal cables)
c) The bar grid, developed into
a plane
plane Square net)
Fig. 11 The structure when developed into a plane is a right-angled, square grid of bars of
equal length.
20 CONCEPTUAL DESIGN However, this basic quadrangular mesh patter characteristics of a shell to withstand
20
CONCEPTUAL DESIGN
However, this basic quadrangular mesh patter
characteristics of a shell to withstand wind an
braced diagonally with thin cables to achieve
vary
in length, entailing the never ending task
of
bars, cables are used, running beneath the
clamping plates at the joints. Therefore, there
ging length of the diagonals. Later the cables
the entire length without any problems. The p
and compression, and very efficiently support
onto the flat bars. Consequently the glass pan
angles (fig. 12).
^*>^
>v.
^v
K*
"
J. SCHLAICH, R.
J. SCHLAICH, R.

ttiffittltv

Fig. 13 Three-dimensional graph and interior view

22 CONCEPTUAL DESIGN - J&B Fig. 14 Aerial view of the completed roof stressed against
22
CONCEPTUAL DESIGN
-
J&B
Fig. 14 Aerial view of the completed roof
stressed against each other by the suspender

under sufficient prestress (fig. 15). This primar

ments, is sufficiently stabilized in itself and ne

structure. Horizontal wind forces are transferre between some support columns, arranged in th

The 40 cable girders are made by the upper "s

J. SCHLAICH, R. < ^ 3 ~«~«. "i Fig. 1 6 Installation of membrane panels
J. SCHLAICH, R.
<
^
3
~«~«.
"i
Fig. 1 6 Installation of membrane panels

Tubulär steel arches with a tie span the distance

nodes in circumferential direction. Seven such ar

producing 6 saddle-shaped membrane roof parts running frorn the last arch to the edge cables (fig

jmmmm'M

#k 24 CONCEPTUAL DESIGN OF LONG-SPAN ROOFS 4. Sobek, W.: Auf pneumatisch gestützten Schalungen hergestellte
#k
24 CONCEPTUAL DESIGN OF LONG-SPAN ROOFS
4.
Sobek, W.: Auf pneumatisch gestützten Schalungen hergestellte
Betonschalen. Dissertation Universität Stuttgart, 1987
Schlaich, J., Sobek, W.: Suitable shell shapes. Concrete International Jan. 1986
Schlaich, J.: Do concrete Shells have a future? lASS-Bulletin Nr. 89, June 1986
5. Nooshin, H., Makowski, Z.S. editors: International Journal of
Space Structures. Multi-Science Publishing Co. Ltd.
6. Leonhardt, F., Schlaich, J. et al: Vorgespannte Seilkonstruktionen - Das Olympiadach in
München. Der Stahlbau Sept., Oct., Dec. 1972, Febr., March, April, June 1973
7. Schlaich, J., Mayr, G., Weber, P., Jasch, E.: Der Seilnetzkühlturm Schmehausen.
Der Bauingenieur, Nov. 1976
8. Bergermann, R., Schlaich, J.: Cable-Membrane Roof for the Arena in Zaragossa, Spain
Structural Engineering International 4/92
9. Trofimov, V.l., Eremejev, P.G.: Thin Sheet Metal (Membrane) Roof Structures.
IASS - Madrid 1989
10. Schlaich, J.: Direct Use of Solar Energy. Structural Engineering International 2/94
11.
Schlaich, J., Schober, H.: Glass-covered Lightweight Spatial Structures. IASS-ASCE
International Symposium 1994, Atlanta, Georgia, USA
25 Technical/Economic Evaluation of Cables for Long-Span Structures Reflexion technico-economique sur les cables de
25
Technical/Economic Evaluation of Cables for Long-Span Structures
Reflexion technico-economique sur les cables de structures de grande
portee
Technisch-ökonomische Betrachtung von Seilen für weitgespannte
Tragwerke
Michael J. COOK
Edmund HAPPOLD
W. lan LIDDELL
Group Director
Chairman
Manager
Büro Happold
Büro Happold
Büro Happold
Bath, UK
Bath, UK
Bath, UK
John W.S. HEARLE
Roger E. HOBBS
Michael R. PARSEY
Chairman
Director
Managing Director
TTILtd
TTILtd
TTILtd
London, UK
London, UK
London, UK
SUMMARY
Large 'tents' and
'airhouses', which have achieved the Status of permanent
structures,
require ropes
or cables to
support
loads. Collaboration between materials
scientists and
engineers
is
essential for
advance. Compatibility with cladding
membranes and other
factors
leads to a list of engineering requirements. Modern
exist in a large number
ropes
of materials, particularly high-performance fibres, and constructions, whose quasi-static
and long-term properties must be selected to meet the needs.
RESUME
Les cables ou faisceaux de cable sont indispensables au transfert des charges
des
"tentes" et des "bulles" de construction permanente. Les
dependent d'une
etroite
progres
collaboration entre ingenieurs et specialistes de la science
des
materiaux. Les exigences
techniques requises decoulent de la compatibilite des cables avec les membranes de
couverture, tout comme d'autres facteurs. Les cables modernes sont realises dans une
grande variete de materiaux, surtout les fibres ultra performantes. Le
etre choisi en fonction des proprietes quasi statiques et de longue duree
type de cable doit
requises.
ZUSAMMENFASSUNG
'Zelte' und Traglufthallen, die den Rang permanenter Bauten erreicht haben, benötigen
Seile oder Drahtbündel zum Lastabtrag. Ein Fortschritt kann nur in
enger Zusammenar¬
beit von Materialwissenschaftlern und Ingenieuren erzielt werden. Die
Verträglichkeit
mit
Verkleidungsmembranen und anderen Faktoren ergeben eine Liste entwurfstechnischer
Anforderungen. Moderne Seile gibt
es in vielen unterschiedlichen Werkstoffen - insbe¬
sondere hochleistungsfähigen
Fasern- und Aufbauarten, deren quasi-statische und
Langzeit-Eigenschaften den
Bedürfnissen entsprechend ausgewählt werden müssen.
J#k 26 TECHNICAL/ECONOMIC EVALUATION 1. INTRODUCTION Since civilisation began, man's built environment has
J#k
26
TECHNICAL/ECONOMIC
EVALUATION
1.
INTRODUCTION
Since
civilisation
began,
man's built environment has been defined more
by
the
materials
available
than
by
his
imagination.
Where there was
nothing
but
ice,
man developed
the
igloo;
where there
was
mud,
man produced
bricks;
where
there
were
fibres
and
skins,
man developed the
"tent".
In
tents,
the
interior
volume
is
enclosed
by
a
membrane,
usually
a
textile
fabric,
whose
tension
carries
its
own
weight
and
any
applied
loads
(due
to
snow,
wind,
etc)
The membrane
is
held
up
in
the
air,
supported
by
rigid
elements
in
compression
or,
less
efficiently,
bending,
or,
in
the
more
recent
development of
"airhouses",
by
air
pressure.
In
the 20th
Century,
polymer
engineering
led to
high
strength
fibres,
weather
resistant
plastic
coatings,
and
high
strength
films
with
life-time
Performances
well
in excess of
earlier
natural
products.
This has allowed
tension structures
to
develop
rapidly
in
recent
years
[1],
achieving
permanent
structure Status
in
several
countries,
including
the
USA.
However,
the
materials
are
not
yet
perfect,
and
nor are
they
necessarily
being
used
in the most
effective
way.
Engineers
have found
design
and
manufacturing
techniques
which
allow
currently
available materials to
be
used
safely
and
reasonably
economically.
But,
as
materials
science
advances,
the
structural
engineer
must
also
move
and
influence
it
into
directions
where
there
is
a perceived
need
or
exploitable
opportunity.
Collaboration
is
vital.
In
this
cables, though
this
cannot be
paper,
separated frorn the developments
we
concentrate
on the
role
of
ropes
and
in
membranes.
In
principle,
in
simple
tents,
such as the classical bell-tent and modern frame
tents,
and
in
airhouses,
ropes
are
not
needed,
except
perhaps
as
guy-lines
to
hold
out walls
or
to
give
added
stability:
the membrane
itself
fills
all
the
mechanical needs.
In
other forms
of
tent,
masts have to
be stayed
by
lines.
However,
in
large
structures,
there is
a
need to
gather the
surface
tensions in
the
membrane
into
linear
elements,
such as
cables,
ropes,
and
webbings,
before
transferring
the tensions
to
the
supporting
medium.
This
requirement
has been
intensified
by the introduction of
new
light-weight
films,
which
have no
fabric
reinforcement.
To control
strain,
they
can
only
be used
in
conjunction
with
linear
elements of
stiffer materials
acting
in
tension,
compression
or
bending.
A total structure will consist of a
number
of
fields
bounded
by
the linear
elements
(cables,
beams
etc).
Each
field may be
composed of
a number
of
fabric
panels sewn or
welded together.
In addition
to
their
principal
structural
functions,
linear tension elements
have many
other
roles
in
places
of
assembly
and
long-span
structures.
These
ränge
frorn
demanding
purposes,
such
as
the
Support
of
movable equipment,
to
minor applications,
such
as
cords used to pull
curtains.
2.
REQUIREMENTS
FOR LINEAR
STRUCTURAL
ELEMENTS
The
scale of structures
using
stressed membrane skins
to
carry
the load
is
limited
by
the
strength
of
available membranes.
To control
deflections in
a
tension
structure
the
membranes must be
prestressed.
This can be achieved with
flat
fabric
fields,
but
greater
stiffness
can
be
gained
by
introducing
anticlastic
curvature
(saddle
shapes)
into
the surface
to
produce
immediate
geometric
stiffness
under
applied
loads.
For
large
structures,
the size
of
a
membrane
field
is
usually
limited
by
the
strength
of
the
fabric
and
the degree
of curvature
built
in.
To achieve
sufficiently
high
curvatures,
and
so
limit
stresses,
the
surface
of
a large
structure
must
be
broken
down
into fields
each
with
their
own
anticlastic
curvature,
separated
by
linear tension
elements
(cables, ropes or
belts),
which
collect
load
frorn
adjacent
fields.
Alternatively,
the fabric must be
supported
by
an
independent network of
cables,
which
carries
the
load out of
the
cladding
membrane
directly.
Elements used
as
linear
restraints at the
junctions
of
membrane
fields
(scallops,
ridges
and
Valleys)
are
more
dependent
on
compatibility with
the membrane
Performance than
complete
load carrying
networks.
The
prime
criteria for selection
are:
(1)
Facility
to transfer
load frorn
membrane
to
element,
allowing
for
parallel
and
perpendicular
force
components;
(2)
If
not
free
to
slide,
element
strain
compatible
with
membrane
strain;
(3)
Elements
flexible
enough
to
follow
curves
in
three
dimensions;
(4)
Elements
with
sufficient
stiffness
to ensure load
transfer;
(5)
Consistent,
predictable
in-line in-line
stiffness,
unchanging with
time,
preferably
linear;
(6)
Durability
to exposure
conditions;
(7)
Easy
termination
and
economical
transfer
of
tension
to
other
components;
(8)
Cost,
including
site
handling,
compatible
with
total
structural
cost;
(9)
Required
strength;
(10)
Easy
transportation;
(11) For networks,
easy
linking at
crossing
nodes.
27 M.J. COOK, E. HAPPOLD, W.l. LIDDELL, J.W.S. HEARLE, R.E. HOBBS, M.R. PARSEY In order
27
M.J. COOK, E. HAPPOLD, W.l. LIDDELL, J.W.S. HEARLE, R.E. HOBBS, M.R. PARSEY
In
order to understand the nature of the forces on and
deformation of typical
linear
elements
in
a tension structure, we consider two examples.
(1)
A
typical
structure,
spanning
40m and
60m
long,
consists
of two
central
fields
and
two
end
fields.
Maximum fabric Stresses
in
the
central
fields
(due
to
applied
loads)
will
be
about
25kN/m,
creating
boundary
cable
tension
of
200kN
and
ridge
cable
tension of 500kN.
A
common
difficulty
arises at the
field
boundaries
occur in the
due to the
anisotropy
of the material.
The
highest
stress
levels
warp
or weft
yarns,
which are
oriented in the
direction
of
principal
curvatures.
Since
the
yarns
are
rarely
orthogonal
to
the
edges
of
fields
there
are
both
perpendicular
and
tangential
forces
on
the
linear
elements
at the
boundary.
Usually
the
tangential
forces
are
transferred
to
the
linear
elements
through
sewn-on
edge
belts
or
bolted metal
edge
clamps,
thus
transforming
the shear forces
into
rope
tensions.
The
normal
forces
are
directly
carried
by
sideways pressures
on the cables
or
ropes,
with the
residual
force
balanced
by
components
of
tension
in
a
rope
following
a
curved
path.
Smooth transfer
of
these
forces
is
difficult
to
achieve,
especially
where
fabric,
belt
and cable each have
their
own
nonlinear
elasticity,
and where
connections
are
imperfect.
This
Situation
would be
simplified
if
one element
were
suitable
for
carrying all
the load
frorn the fabric
regardless
of
weave
orientation.
(2)
In
a
typical
airhouse,
the connection between fabric
and ground
perimeter
demonstrates
the
problems
created
by
incompatible
straining.
The
fabric
strains
under
in-plane
longitudinal
Stresses,
but
the
ground
will
not strain.
This
results
in
shear
distortion of
the
membrane towards
the
ground.
An
airhouse
spanning
more than
40m
will
require
cable
restraint
to
limit
in-plane
fabric
Stresses,
and
define distinct
areas
of membrane to
transfer
load
into
the
network.
The form which
they
take
must be
predicted,
so
that the
fabric can
up
be
patterned
to match
it.
Hence
predictability
of
strain
is
a
very
important
feature.
Compatibility
of strain
becomes
an
issue when
the
restraining
cables
are
not
free
to
move
across the
membrane
surfaces.
This
effect
will
occur where
two-directional
cable
nets are
used to "lock-in"
to the deformed
membrane
surface.
Design
loads
may be over 500kN for cables in
large
roofs,
but actual
loads
are
highly
variable.
Usually
the
prestress
levels
will
be
only
10 to
20%
of the
ultimate
design
load,
which is
predicted
to occur
only
once
in
the
building*
s
lifetime.
However,
the
frequency
and
magnitude
of
load
vary
with
the
patterns
of
wind and
snow
forces.
Wind
loads,
having
a
significant
dynamic component,
will
create
short term
peaks
of
stress
measured
in
seconds,
whilst
snow
loads
may have
durations of several
days.
Whether wind
or
snow
is
the
critical
load
case
will
clearly
depend
on the
location of the structure.
In
addition
to the
dominant
in-line
tension,
any
curvature leads
to
tensile and compressive
bending
strains and
other
forces: these are greater at
detail
points
such
as
eyelets.
For membranes where
tear
propagation
is
the
dominant failure
mode,
design
Stresses must be
sufficiently
limited
to ensure that
relatively
small
tears
remain stable.
Factors
of
safety
between
8
and
6
on
tensile
strength
are
commonplace.
For
linear
elements,
factors of between 2 and
3,
on
ultimate
breaking
strength,
are
normally
acceptable
when used as
an
integral
part
of
a
permanent
structure,
whilst
a
factor
of
5
would be used
for
stand-alone
lifting
equipment where
the
Chance
of overload
is greater and the handling might
cause
damage.
In
building,
construction at least
cost
is
a
vital
consideration.
Investment
decisions
normally
underemphasise
"least
lifetime
cost"
in
favour of
"least
initial
cost".
Membranes
currently
used
are
PVC coated
polyesters
with
lifespans
of
10
15 years
and
PTFE
or
silicone
coated
glass
fibre
with
lifespans
of
25
to to
35
years.
The additional
linear
elements
must have
at least
a comparable
lifespan.
If
a
replacement
membrane
is
planned,
longer-life
elements
might
be
appropriate,
but
there is
a
strong
possibility
that
new
materials
with
better
characteristics
and
different
properties
will
become
available,
and
so make
complete
replacement more
logical.
Where
the membrane
merely
clads a cable
network,
and
if
strain
compatibility
is
not
important
(as
with
a
polymer
panel
cladding
on
a
cable
net),
independent
replacement
of
the
cladding
is
highly
feasible.
Therefore there
be
sound
economic
may
justification
for
selecting ropes or cables that
outlive the membrane.
3.
MODERN
FIBRE
ROPES
Frorn ancient
times until 150
years
ago,
ropes
were made from natural
fibres;
then steel wire
ropes
displaced
fibre
ropes
from serious
uses
in
structural
engineering.
The
Situation
changed again
50 years ago with
the invention
of two
28 TECHNICAL/ECONOMIC EVALUATION generations of strong man-made fibres, the first (nylon, polyester etc) more
28
TECHNICAL/ECONOMIC
EVALUATION
generations of
strong
man-made
fibres,
the
first
(nylon,
polyester etc)
more
extensible,
and
the
second
(carbon,
aramid
etc)
with
high
modulus.
Since these
were
continuous filaments,
high
twist was not
essential,
and
ropemakers
have
developed
of
constructions.
The choice
is enormous:
it
a ränge
new
is
estimated
that,
for
any
given
purpose,
500,000
combinations could be considered.
Architects
need
specialist
advice
from experts who know the
field
and have
facilities
for
Computing
Performance.
In
this
paper,
we can
only
give
some
general
Information,
more
related
to the main
structural
need
formulated above
than
to the
anciiiary uses also mentioned in the
introduction.
Table 1 lists
materials
currently
available for
ropes,
with
rough
indications of
strength
and
stiffness
relative
to
weight
and
price;
more detailed
plots
have
been
given
elsewhere
[2].
Except
for
low cost
for
minor
uses,
which
favours
polypropylene,
and
anciiiary
uses
needing
high
extensibility
and energy
absorption,
which favour
nylon,
the
choice
is
between
polyester,
at
lower
cost,
or
the
high-modulus
high-tenacity
(HM-HT)
fibres
at
lower
weight.
The
brittle
(in
bending)
fibres,
carbon
and
glass,
are
only
suitable
for
use
when bonded
by
a
matrix
as
solid
pultruded
rods.
Such rods
also
be used with other fibres,
may
in
order to reduce
the
problems of axial compression.
natural
fibres with limited
Performance
hemp,
sisal,
cotton etc
[1/0.1/2/2]
man-made fibres
adequate
for
common
uses
regulär melt-spun
Polyethylene, polypropylene [2/0.2/5/1]
intermediate Performance
fibres
polyamide
(nylons)
[3/0.5/4/1]; polyester [3/1/4/2]
glass
[5/1.5/10/10]
advanced
[HM-HT]
fibres from
liquid
crystals
aramid (®Kevlar
and others); LCAP
(®Vectran); PBO [7/4/1/1]
other
high
modulus,
high
strength
(HM-HT)
fibres
high-modulus
Polyethylene
(HMPE)
(®Spectra,
®Dyneema)
[10/7/1/1.5];
carbon
[8/10/0.4/2]
steel wire [1/1.5/1/5]
Table 1 Materials with
approximate
ratings
for:
[strength-5-wt/stif fness^-wt/strength^-price/stif fness-5-price]
Ropes
and cables
be
categorised
in
five broad groups.
Twisted constructions
may
comprise
the
familiär three
and
four
Strand
ropes.
High
twist was
necessary
to
hold
short
fibres
together, but
the
tensile
efficiency
is
low and
torque
balance
is
poor.
Ease
of
splicing
on
site
leads
to their
use
as guy
lines
for
tents
and
marquees.
Plaited
and
braided
constructions,
interwoven
from
equal
numbers
of
clockwise or counterclockwise
Strands,
be either solid
may
(e.g.
8-strand)
or
circular.
Single
circular
braids
have
hollow centres.
Double
braids
are
made
with
a second
Single
braid
over
an inner braid.
Stranded
(wire
rope)
constructions
are
similar to
twisted
constructions,
but have
lower
twist
to
improve
tensile
Performance.
The Strands
(or
wires)
are
arranged
in
concentric
rings,
and
may
be
designed
to
minimise torque
generation.
They
should be used
when
the
application
requires
the
rope
to
be
worked
over
sheaves
(pulleys)
at
high
loads.
Whereas
jacketing
is
merely
desirable
in
some other constructions
to
provide
wear
and
light
resistance,
it
is
essential
in
stranded
rope to hold
the
construction
together.
Parallel
constructions comprise
parallel
yarns,
Strands
or sub-ropes
held
together by
an external
cover,
which
extruded
may
be
polymer
or
a
braided
jacket.
They
have
high
tensile
efficiency,
but should not
be
worked over sheaves
at
high
loads.
Some
types
need
special
terminations.
Pultrusions,
which
have
excellent
strength and
fatigue
resistance,
but
are
costly
to
make
and
difficult
to
terminate,
comprise
an
assembly
of
fibres
in
a
rigid
or
flexible resin.
Other
linear
elements
include
flat
woven or braided
webbings,
which
be easier
may
to join
to
membranes
or to
use
as
slings.
Round
slings,
consisting
of
parallel
yarns
in
a textile
casing,
offer
much
higher
specific
strength
for
lifting
applications.
Since
variants of the
many
above
types,
including
blended
forms,
are
possible,
it
is
easy to see
why
the choice
is
so
large.
Furthermore,
terminations
cause as
problems
as the
may
many
ropes
themselves;
the
available
types are
variants
of
grips,
splices,
resin
sockets,
and
barrel-and-spike.
Computer
programs
have been
developed to predict
the
load-elongation
and torque-
twist
properties
of ropes
[3].
These
can
be
used to design
ropes
with
M.J. COOK, E. HAPPOLD, W.l. LIDDELL, J.W.S. HEARLE, R.E. HOBBS, M.R. PARSEY 29 properties which
M.J. COOK, E. HAPPOLD, W.l. LIDDELL, J.W.S. HEARLE, R.E. HOBBS, M.R. PARSEY
29
properties
which match the membrane
properties,
and meet the other
requirements
specified
above.
By
of
way
example,
Table
2
gives properties of a few
candidate
ropes
for the strength
requirement
of
500kN.
FIBRE
ROPE-TYPE
WEIGHT
STIFFNESS
PRICE
NOTES
kg/lOOm
kN/1%
extn
£/100m
manila
three-strand
500
40
600
GRADE
1
steel
galvanised
300
360
700
6X36IWRC
plastic
coated
1000
stranded
stainless
2500
poly¬
8-strand
200
20
500
worked &
propylene
X-plait
rested
polyester double
braid
220
30
1500
ditto
nylon
double braid
150
15
1200
ditto
polyester parallel
150
40
1000
ditto
Strand
carbon
pultrusion
70
600
10,000
1x6 helical
aramid
parallel
50
180
2500
w'rkd &
r'std
yarn
Kevlar
29
HMPE
stranded
40
80
1600
stiffness 0.1Hz
Table 2
Rope
properties at 500kN strength (approximate values only), excluding
terminations.
In addition
to the
quasi-static
properties,
long-term
Performance is vital.
Failure
often
occurs
at
terminations.
Of
the
six
"fatigue"
mechanisms
identified
in
ropes
[4],
the
following
are
likely
to
be
of
most
importance in
building
structures.
Creep
rupture,
determined
by
"average peak
load",
is
a
major
weakness
of
HMPE,
and,
to
a
lesser
extent,
nylon,
and
is
always
the
default mechanism
in
the
absence of others.
Internal
abrasion
is
most serious
in
nylon.
Axial
compression
fatigue
is
a
potential
killer
in
HM-HT
ropes,
and,
to
a
much lesser extent
in
polyester
and nylon.
Even
when
the ropes
themselves
are
always
under
positive
tension,
axial
compression
can occur
(and
has caused
embarrassing failures
in
marine
uses),
especially
with
low
minimum
loads, due to
twisting
of
ropes
that
are not
torque-balanced,
differential
response
of
components,
or
bending.
Computer
programs
have been
developed
to
model these
fatigue
responses
[5]
(and
hysteresis
heating),
though
the
theory
is less
certain
than
for
the
quasi-static
behaviour
and
input
data
are
not always
available.
4.
THE COMPETITIVE
POSITION
At
present
engineers
are
generally
selecting
steel
ropes and Strands wherever
high
strength,
high
stiffness
and
durability
is
needed.
Steel dominates the
field
for
structural
rigging,
"cable"
nets
and
boundary scallops.
However,
steel
is
not
ideal
for
the
following
reasons:
steel
ropes
have
a
nonlinear
stress-strain
relation
at
low
(working)
Stresses;
it
is
difficult
to
transfer
forces
tangentially
into
a steel
cable;
steel
cables
cannot
be
sewn
or
welded
onto coated
fabrics
and
require
special,
costly
additional
clamps;
steel
ropes
can
damage
and chafe
the
fabrics
used;
normal
oils
present
in
cables
can
härm
fabric
coatings;
for
transport,
steel
cables need
to
be
removed
from the fabric
panel
to
allow
it
to
be
packaged
small and to avoid
damage;
steel
cables and
their
terminations
must
have
corrosion
protection
unless
stainless steel
is
used.
Against
this,
steel cables have a
cost
advantage over
fibres
any
man-made
that
might
be
able
to
perform
better.
Advanced
fibres,
such as
Kevlar,
offer
potential
advantages
of
high
and
reliable
stiffness
and
light
weight.
They are
attractive
to
the
structural
engineer
for
these
reasons,
but
their
use can
rarely
be
justified.
Their
most
notable
use
is
as
stiffeners
to
sails
for
racing
yachts,
where
relatively
small
improvements
in
Performance and
control
win
the
race.
In building
structures,
properties
are
rarely
that
critical
and
price
dominates.
However,
with
current
advances
be the case
in
the
technology,
the
cost
advantage
of steel
not continue to
for
may
long,
especially
when the
indirect
cost savings
as well
as the
direct cost are taken
into
account.
30 TECHNICAL/ECONOMIC EVALUATION The clearest advantage of the most recently developed fibre ropes over steel
30
TECHNICAL/ECONOMIC
EVALUATION
The clearest
advantage
of the most
recently
developed
fibre
ropes
over steel
cables
is
low
weight-to-strength
ratio.
This
can
be
useful
for
mobile
structures to
give
ease
of
handling,
but becomes
especially
relevant
for
extremely
large
air-supported
structures
covering
many
hectares.
For
a
structure
spanning
500
meters,
the
weight
of
restraining steel
cables
is
likely
to be between
8
and
15kg/m2.
With normal
inflation
pressure
of
20
to
30 g/m2,
such a
weight is
very
significant;
therefore
operating
pressures
would
have to
be increased.
However,
recent
research shows
that
the
sensitivity
of
such
structures to
aerodynamic instability,
as well
as
buffeting
from
the
wind,
depends
on the
net
over-pressure.
Using
an aramid
rope
of
the
same
strength,
the
weight
of the
restraining
cables
would reduce to
between 2 and 4kg/m2.
In
addition
to the
stability
advantage,
the
saving
in
inflation
equipment,
easing
of
access route
pressurisation
problems,
and
reduction
in
required
strength
of
cladding
membrane
needed
to
withstand
the
pressure
could be
sufficient
to
pay
for
the
additional
capital
cost of the cables.
Design
studies
would be needed
to determine the
long-term
durability
of
such ropes,
particularly in terms of
the need to avoid
axial
compression
fatigue.
The
above example is based on a technical Solution in terms of
rope
strength;
but there
is
also
a
strong
case
for
considering
an intermediate
Solution
using
polyester
ropes.
The
weight
would be
greater,
thus
reducing
some
of
the
benefits,
but
the
rope
cost
would be lower
and current
indications
are that
polyester
is
more
rugged
in use.
Rope
experts
have
more
confidence
in
its
resistance
to
fatigue.
Another consideration
is the
axial stiffness
of
the
rope,
which needs
to
be considered in
relation
to compatibility
with membrane
properties.
5.
CONCLUSION
It has taken some
forty
years
to
develop
mebrane materials which can
adequately
carry
loads and which
have
a
decent
life.
Simultaneously,
it
has been
necessary
to
develop
analytical
tools
for
engineers
who
design
with
these materials.
Now
that
both
are
available,
architects
world-wide
are
designing
tents.
In
large
structures,
the
membranes
are bordered and
prestressed
by
linear
tension
elements,
which
not
only
gather
high
forces
but
act
as
tear Stoppers.
These
elements
have
their own
problems
and,
again,
the
role of
new
materials must be
addressed.
As
well
as the
intrinsic
properties
of
ropes,
cables and
webbings,
questions of connections,
including
clamps
and
terminations, come up.
Force
is
not the
principal
issue.
It
is
more
important
to
consider comp^' Lbxlty
-
of
strain
and other
aspects
of physical
behaviour,
Performance in
fire,
lifetime,
and
so on.
Above
all,
is
a
10%
increase in
economy
a
major
factor:
load-carrying
capacity means
little,
but
a
10%
reduction
in cost
is
a
revolution.
This
paper
architects and
encourages
engineers
to
problems
of
linear
tension
elements with an
open
mind,
taking
address the
account of the
availability
of
new materials and
structures
and the
complex
and
many
interacting
features of Performance and cost.
Since the
next
major
building
type,
once
the Visual conservatism and reluctance
to think of
buildings
as
machines
is
overcome,
is
the
airhouse,
which
is
the
most economic
climate
moderator one can
think
of,
the
opportunities
for
new linear tension elements
will
be even more important.
REFERENCES
1. HAPPOLD
E.
and LIDDELL
W.I., Tension Structures, 14th Congress of IABSE,
New Delhi, 1992,
137-150.
2. HEARLE
J.W.S.,
BURGOYNE
C.J.
and
HOBBS
R.E.,
Fatigue
and
Length
Effects in
Fibre Ropes.
1993.
Workshop,
El Paular, Madrid
1992.
IABSE,
Zürich,
66,
307-318,
3. LEECH
CM.,
HEARLE J.W.S.,
OVERINGTON
M.S.
AND BANFIELD
S.J.,
Modelling
Tension
and
Torque
Properties
of Fibre
Ropes
and Splices,
Proc.
3rd
Int.
Symp. Offshore
and
Polar
Engg.,
Vol.
2,
370-376,
1993.
4. PARSEY
M.R.,
BANFIELD
S.J.
and
HEARLE
J.W.S.,
Life of Polymerie Fibres in
the Marine
Environment.
2nd
Int.
Conf.
on
Polymers
in
a Marine
Environment,
Inst. Marine Engineers, London,
1987.
5. HEARLE J.W.S.,
PARSEY
M.R.,
OVERINGTON
M.S.
and BANFIELD
S.J.,
Modelling
the
Long-Term
Fatigue
Performance of
Fibre Ropes.
Proc.
3rd
Int.
Symp.
Offshore
and
Polar
Engg., Vol.
2,
377-383,
1993.
31 The Valeria of Roman Amphitheatres: The Colosseum Les velariums des amphitheatres romains: le Colisee
31
The Valeria of Roman Amphitheatres: The Colosseum
Les velariums des amphitheatres romains: le Colisee
Die Sonnensegel römischer Amphitheater
Giorgio CROCI
Piero D'ASDIA
Dina D'AYALA
Prof.
Researcher
Univ. "La Sapienza"
Assoc. Prof.
Univ. "La Sapienza"
Univ. "La Sapienza"
Rome, Italy
Rome, Italy
Rome, Italy
Piero MEOGROSSI
Architect
Soprintendenza Archeologica
Rome, Italy
SUMMARY
The
paper
reviews the
current archaeological and structural assessments on the layout of
the velarium. The aim is to
suggest
a methodology to understand the architectonic and
structural choices, validating the various hypotheses through the structural
analysis. The
results obtained in this first step
of the research will be used to narrow the
field
of the
possibilities and concentrate
further studies of better accuracy, on the modeis that look
more reliable.
RESUME
Les auteurs donnent une vue d'ensemble des
evaluations archeologiques et structurales
sur les traces des velaires de l'antiquite. II
s'agit
de
proposer
une
methode pour
comprendre les choix architecturaux et structuraux, en verifiant
les
diverses
hypotheses
par
des calculs statiques. La premiere phase de recherche fournit des resultats qui
servent d'une part ä reduire les variantes possibles et d'autre
part
ä preciser les etudes
supplementaires, pour que les modeles presentent une meilleure
fiabilite.
ZUSAMMENFASSUNG
Der Beitrag gibt einen Ueberblick über die gegenwärtige archäologische
Rekonstruktion von Sonnensegeln in der Antike. Ziel ist die Erarbeitung
und bauliche
einer Methodik
zum Verständnis der Entwurfskonzepte, für die unterschiedliche Hypothesen durch stati¬
sche Berechnungen überprüft
werden. Die im ersten Schritt erhaltenen Forschungser¬
gebnisse dienen der
Eingrenzung der Möglichkeiten und der Konzentration vertiefender
Studien auf vielversprechendere Modelle.
32 VALERIA OF ROMAN AMPHITHEATRES 1. The shape of the velarium: topographical and archaeological hypotheses
32
VALERIA OF
ROMAN
AMPHITHEATRES
1. The shape of the velarium: topographical and archaeological hypotheses
The architecture of the Coliseum in specie ovi, as it is called by Cassiodoro, was highly influenced by the
existence on its building site of the oldcr stagna Neronis, an artificial lake, in the Domus Aurea, placed at the
centre of the city in the Valley among the Celium, Palatinum, Velia and Oppium Hills.
The geometry of the plan is an oval with 4 centres of curvatures organised in 80 arch-bays repeated three
times in the vertical and horizontal planes and connected with radial walls which bear the cavea's seats.
The two principal axes, of which the major one is aligned along the Via Sacra, measure 188 m and 156 m.
Only the outer 50m where actually built over the ground with a degrading section 50 m tall at the top of the
actic level of the outer wall. So that the area to be covered by the velarium colore coeli (Pliny, 9, 1, 24)
measures approximately 23000 m2.
The centre of the Coliseum is based on a Pythagorean triangle superimposed to the two axes and the centre
of the valley whose measure is recognised by a specific topographical relation which organise the over all
shape of the monument by an urban axis (fig. 1). This axis links the ara maxima Herculis, to the Curiae
Veteres, two important knots of the Roma quadrata grid narrated by Tacitus, and it connects a number of
significant centres of monuments, from different roman periods, as the Apollum templum by Augustus, the
Flavi's labyrinth, the Adrian's Adonea, Constantinus's Arch, the Nero's octagonal room. This axis, called as
the urbis axis, measures 19° N-NE (corresponding astronomically to the course of the sun of the XXI April,
day assumed to be the Rome's birthday) and it is orthogonal with the ancient via Lata.
Such important considerations support the in situ exam over the moulding and the exam of seven
architectonic pieces erratically dislocated on the archaeological site. They allow to formulate the hypothesis
that these blocks of the same shape and two different sets
of dimensions
(average
measures
being
950x600x500 mm and 720x440x500 mm) were utilised as a further prop of the poles which hold the cables.
The two different dimensions can suggest the idea that the poles had different functions. In particular it
seems acceptable that the bays of the velarium (slack as they look in the Pompeii's fresco), were of the same
extent as the lower structure, so to have a tent, in the shape of a Latin sail, covering each arch bay. The
brackets on the outer side of the acticum level being three for each bay
(fig. 2
there were three poles for
each sail. We therefore think that the central pole was the one to which the principal cables, called rudentes,
80, one for each bay, were connected, and the only one working during the phase of mounting. Once the
velarium had taken shape and place the forces were redistributed on the other two poles by means of
secondary ropes.
The length of the poles is estimated to be about 12 m of which 7 to 7.5m above the moulding of travertine
showing for each bay three rectangular holes of 400x500 mm in which the poles were slipped through and,
resting on travertine brackets spanned 2.25 m., were fixed to the wall by a planking.
The geometry and structure of the velarium therefore seem to follow the same logic as the geometry and
structure of the travertine construction below, conveying, through the repartition of the forces among the
three poles, the flux of Stresses toward the arch-pillars system.
To lay-out the cables net, mounting it and lifting it in the final position took approximately four days and the
work of 300 men of the imperial fleet.
2. The shape of the velarium: structural hypotheses
The structural elements which can be still seen today on the monument and the historical and archaeological
documents, presented in the previous paragraph, lead to convincing hypotheses about the construction of the
velarium and its assemblage and rising in place; however they are not sufflcient to deflne completely the
structure and the building technology used. The structural analysis can be usefully applied to enlighten these
aspects and therefore to reduce the uncertainty associated to different lay-outs
pointing out
among these,
only the ones which fulfil the stress requirements according to materials and technology available at the time.
Confronting the different configurations proposed by various authors, it seems possible to deflne a number of
points which form a common core for further discussion:
a) the main cables present a radial lay-out, one for each bay of the outer travertine arched structure, and they
were probably made of hemp, this being the material commonly used by sailors. Most likely diameters were
M\l G. CROCI, P. D'ASDIA, D. D'AYALA, P. MEOGROSSI 33 in the ränge from 40
M\l
G. CROCI, P. D'ASDIA, D. D'AYALA, P. MEOGROSSI
33
in the ränge from 40 to 60 mm (the cross section being measured when in tension) and each cable was tensed
by a capstan controlled by two to three men. This hypothesis, better than the one with 240 cables, allows to
have homogeneous bays and even distribution of loads and stresses (if not for the slight Variation connected
with the change in curvature along the oval) therefore avoiding asymmetries not liked by Romans and
making the structures of the velarium congruent with the lower structure.
b) 80 capstans were placed above the roof of the giant portico in front of the central pole of each bay, which
held the pulley of the principal cable; during the mounting phase, the central pole bore all the load of one
bay, and it was probably sustained by two other poles put in a shoring; once the cables were tensed and the
velarium put in place, before removing the shoring, the load on the principal cable was shared onto the two
lateral poles by means of secondary cables (see fig. 3) tensed by winches set into blocks of travertine at
purpose shaped. The presence of the boundary-stones, with four holes, placed at the ground level outside the
monument, has been interpreted by some authors as an anchorage for the cables. It seems unlikely for a
number of reasons: a huge increase in the amount of materials, a difflculty in the coordination of the
Operations due to lack of visual connection, the hindrance that these outer cables could cause during the flux
and defluxion of the audience.
c) the canvases were directly tied to the principal cables in roughly rectangular pieces of width slightly
longer than the distance between the two adjacent cables and length of 3 to 4 m., so to have elements of
workable dimensions (20 to 30 m2 was the average dimension of Latin sails) and to allow independent
movements of the different pieces under the action of the wind, so to reduce the global pressure;
d) the inner ring to which are connected the low points of the principal cables, repeats, with smaller
dimensions, the shape of the oval plan of the structure, sharing the same centres of curvature. The four arcs
in which the oval is divided by the change of curvature have the same length
but different angles, the one
with smaller radius being 1.84 rad, the other 1.30 rad.
Considering the thrust of the cables (20 on each
are
equally spanned) like a pression uniformly distributed over these arcs, one obtains an hoop thrust which
differs with respect to the mean value (calculated for even angles of;r/2 rad) for a ± 17% As it is no
possible in such a structure to have variable thrust this means that the mean value is reached by means of a
change in sag in the cables: for the cables in the area with smaller radius of curvature the sag will be reduced
of 17% and on the other it will be increased of 17% so to have an opposite increase and reduction of the
hoop thrust in each area. The inner ring will show a saddle shape similar to many roofs of modern Stadiums.
Starting from these points of common agreement a number of questions remains open:
1- the actual conflguration of the principal cables, the covered surface, the limitations imposed by the
visibility line and the possibility of shadowing the podium related to the position of the sun;
2- the connection of the canvases to the cables: the possibility of moving them along the cable means a
further load due to the secondary mesh of rope;
3- the conflguration and realisation of the inner ring made of hemp or iron rods connected by small rings;
4- the actual dimension of the poles, the need of struts during the mounting phase and the compatibility of
the stresses;
5- the action of wind and rain, at least in their mild summer manifestations, and their compatibility with the
global lay-out of the velarium and the dimension of its structural element.
3. Structural analysis
To work out the structural analysis we need to define the mechanical features of the material and the
distribution and entity of actions. The nominal values mentioned below have been assumed from old codes
or from materials which are today similar to the old ones:
cross sect.
weight
failure stress
elastic modulus
hemp cables
2000 mm2
30 N/m
30-40 N/mm2
300-400 N/mm2
oakwood poles
0.12-0.20 m2
1000-2400N/m
40-60 N/mm2
12000 N/mm2
iron bars
900-2500mm2
70-200N/m
200 N/mm2
210000 N/mm2
linenclothes
1. m2
ION.
34 VALERIA OF ROMAN AMPHITHEATRES To verify the first point mentioned above, the lower bound
34
VALERIA OF
ROMAN
AMPHITHEATRES
To verify the first point mentioned above, the lower bound for the sag, not to interfere with the visibility line
of the upper seats, is 25 m. (flg. 4). Although some of the archaeological reconstructions show the velarium
as practically flat, the analysis shows that the Upper bound can be not less than 7 - 11 m for cables spanning
from 40 to 60 m. The minimun. of the last two measures, 40 m., is conditioned by the extension and duration
of the shadowing over the lower seats, while the maximum span is bounded by the increase in load with
relation to the strength of the cable. The analysis proves that for the longer span tested (61 m) is not possible
to have the overioad of two additional cables to movement the canvases and therefore in this way is
implicitly solved the second point of the previous paragraph (flg. 5).
The geometry and cross section of the inner ring is related to the length of the cable on one side and to the
horizontal component of the trust (Th) on the other, assuming that the sag in the cables along the ring will be
variable as already discussed at the point d) of the previous paragraph. The hoop stress in the ring will be:
N 40r
12,77^
and varies
for a sag of 9 m, from
171.5 KN to 381 KN for a span ranging from 40 to
n
60 m, while for a sag of 15 m the corresponding values are 127 KN to 222.5 KN. For the span of 61 m
the
case of reduced load due to unmovable canvases has also been considered, producing a hoop force of 222.5
KN and 159 KN, for sags of 9m and 15m respectively. Taking into account the variability of the distance
between cables at the lower points as a function of their span, an average weight of the inner ring on each
cable of 500 N can be assumed.
According to the span and sag ranges considered, the cable reaches the pole forming an angle with the
horizontal variable between 18° and 29°, while dimensional analysis implies that the position of the capstan,
placed as close as possible to the pole, will give to the cable Coming from the pulley an angle with the
vertical line of approximately 20°. This means that the resultant of the thrust will form with the axis of the
pole an angle of 40° to 46°. Therefore the horizontal and vertical components will be respectively in the
ränge of 1.18 T to 1.30T and 1.4IT to 1.30 T, giving place to the following generalised stresses in the poles:
M=O6.6 + V03,
S
0,
N
V
t
r
A^„,
->
.The stresses for the most severe case vary from 11.3 to 14.23
N/mmz
for cross section of 500x400 mm, and from 24. to 31.6 N/mm2 for cross section of 400x300 mm.
Confronting this values with the ultimate strength value it can be seen that the pole is able to resist those
loads with a safety factor variable between 3.6 and 1.3 depending on the cross section.
4. Conclusions
The results obtained represent the first iteration of a process of trial and error that should be replicated
introducing the Variation suggested by new Information and their critical analysis. However they are useful to
bound the field of the possible Solutions. In particular:
- taking into account that the decrease in load (from 25 to 15m sag) is of about 18 %, while the increase in
thrust is far from the limit value, it can be thought that the more likely shapes were the ones with sag variable
from 7 to 15 meters, depending on the span. It is worth to mention that, when observed from below at a
distance of approximately 50 m from the lowest point, over an overall length of 180 m. entities of the sag
from 10 to 15 m could have given the impression of a flat velarium.
- while for the smaller spans and bigger sag is still possible to think that the ring would be made of hemp
ropes (requiring a section of approximately 15000 mm2 with a weight 265 N/m)
when the longer span are
considered, and always for sags close to the upper limit value, the iron rods appear convenient, involving
much smaller section and therefore a weight approximately estimated in 180 N/m.
-at least during the mounting phase, for the longer spans of the cables and smallest cross section of the poles
a strut was required to lift up the cables net.
It is not possible here to discuss in due depth the reduction of the global pressure of the wind obtained by the
cuts in the canvases, further studies and possibly tests on a model in the wind tunnel being necessary to
achieve reliable knowledge of the phenomena involved. The effects, however, could have been partially
absorbed by relatively small sub vertical stays tied at the lower levels of the seats. The stays would have
G. CROCI, P. D'ASDIA, D. D been slack o slightly tensed, so not to affect
G. CROCI, P. D'ASDIA,
D.
D
been slack o slightly tensed, so not to affect the confl
vertical action due to the rain water or the wind flowing a
above results.
For the rain, the load previously taken
waterproof treatment by linen oil and a water film 0.5 m
and their slope allow to assume that, even during a fairly
on average, greater than 35% or 70% according to the di
and down the canvases. This simple analysis, therefore
security factors mentioned above, suggests that very like
exceeded the 50 to 55 m, and that for such configurations
As for the wind, with a simplified equivalent static analy
speed of the wind will have been
14-15 m/s.
Such
a
sp
force that in the depression area would equal the dead loa
the safety factor to 1.5, which seems not reliable. For the
not of much influence the increase or decrease in the hori
above.
36 VALERIA OF ROMAN AMPHITHEATRES 5 10 1» 20 25 30 » \J*1 L=E^ <3>
36
VALERIA OF ROMAN
AMPHITHEATRES
5
10
20
25
30
»
\J*1
L=E^
<3>
Fig. 3 and 4 : Possible shapes of the velarium
Thrust in the principal cables for various load cond.
4500
4000
3500
3000
2500
47m
load
max.
2000 4Q:ms
load
mm. load
1500
54 m.
1000
10
12
14
16
18
20
sag (m)
Valdek KULBACH Prof. Dr Tallinn Technical Univ. Tallinn, Estonia —:
Valdek KULBACH
Prof. Dr
Tallinn Technical Univ.
Tallinn, Estonia
—:

Design and Erection of Long

Conception et montage de structures ret Entwurf und Montage weitgesp

-.-,

38 ßk DESIGN AND ERECTION OF LONG-SPAN 1 PRELIMINARY REMARKS For many types of long-span
38
ßk
DESIGN AND ERECTION OF LONG-SPAN
1
PRELIMINARY REMARKS
For many types of long-span structures the round or oval layout may be statically most suitable
On the other side, for many assembly and sports buildings that form of building may be also
more functional as common rectangular one For instance, when we have in the center of
building a great sports field with a running track, the round oval plan enables to locate the
spectators' seats according to conditions of the best vi^ibiiny The same is vahd for a number of
other assembly buildings
Especially suitable for that kmd of conditions may be buildings with
the roof,
formed as
a surface
with
negative Gaussian
curvature or so-called saddle-formed
roofs In the number of saddle-formed roof structures very important position have the cable-
networks In the following we shall investigate the most suitable types of suspended roofs,
formed as prestressed cable-networks inside a closed contour beam
2 INITIAL FORM OF NETWORKS
2 1 General remarks
The form of a prestressed network will be determmed by the conditions of equilibnum of its
nodes
In
most
cases
the
network
will
be
formed
by
two
famihes
of
cables
crossing
Theoretically we have a plurahty of possibilities for choise of conflguration of network surface
In every case the Gaussian curvature of the network surface has to have a negative value in all
its nodes The most suitable roof structures for building practice to our mind are orthogonal or
approximately orthogonal networks
In the following we shall investigate mamly the orthogonal
networks
Approximately orthogonal may be considered the networks, formed by free mutually
sliding cables at the time of prestressing the network
In the first case the horizontal
component
of inner force is constant on its length
In the second case invariable on the length of the cable
is
its inner force
itself
In the first case the contact forces
between
carrying
and stretching
cables will be vertical, in the second case they will be applied in the direction of the bisector of
the angle between the neighbourmg sections of cables
For simphfication of our problem we
the network
may
image
as a system of two contmous
famihes of Virtual cables, uniformly distnbuted on the length of
between the nodes In
spacing
this case the cross section areas of cables will be charactenzied by the effective thicknesses of
the famihes of carrying and stretching cables
In this case the condition of equilibnum will be
presented as differential equations The other possibility is to analyse the network as a discrete
system For networks with a great number of cables the real network
may
be replaced by a
Virtual one, which consists of a reduced number of cables, the cross section areas of substituting
cables must be
correspondingly increased
The conditions of equilibnum of nodes in that case
will be algeabnc equations
2 2 Hypar networks
The continous surface of a prestressed orthogonal network may be descnbed by the elliptical
equation
Go|f + "o|f-0
(2
1)
where G0 and H0 - the prestressing forces of the
carrying
and stretching cables correspondingly,
suited to the width unit of the network
V. KULBACH 39 The equation (2.1) will be satisfied for hypar z_/* /3L (2.2) Jxa2
V.
KULBACH
39
The equation (2.1) will be satisfied for hypar
z_/*
/3L
(2.2)
Jxa2
Jyb2
Corresponding prestressing forces will be determined by uniformly distributed contact loads p0
,2
(2.3)
2/,
P^_
(2.4)
Hn
2/v
The most suitable layout for the hypar will be the ellipse (Fig. 1)
x2
v2
(2.5)
a1
b2
where <7, b - semi-axes of the ellipse,
f„fs - the rises of curvature of carrying and stretching cables correspondingly.
^-
,z
.1
y
i
x
X
,1
/1
t
"/
r
/
i
y
-,l
1
x
JD
^rför- *»
r1!
-O
A=B
Fig. 1 Hypar-network
Fig.2 Orthogonal network with round contour
2.3 Orthogonale network with given form of contour
For
the
node
i,k of an orthogonal
weightless
network the condition of equilibrium may be
written in the form
ziMi
zi,k
zq-i
zijc^
zi+\,k
zi,k
^ zi-\jc
zijc ^
)+ "a* (—r
u
n
°
(2.6)
G0,(
L
—:
+ —:
+ —r
where G(H and Hok - the prestressing forces for the /-th carrying and the £-th stretching cable
correspondingly;
alk\ blk - the length protection of the k-i\\ section of the /-th carrying cable and the /-th
section of the k-i\\ stretching cable correspondingly.
For the network with constant mesh dimensions the linear equation (2.6) may be written in the form
_ (\*-i + W + x(Vu + W
(2.7)
\k
2(1 + X)
£k> 40 DESIGN AND ERECTION OF LONG-SPAN where k ——. The system (2.6) consists of
£k>
40
DESIGN AND ERECTION OF LONG-SPAN
where
k ——.
The system (2.6) consists of equations, set up for all internal nodes of the network. The ordinates
of contour nodes have to be given as the initial data of the problem. An orthogonal network,
inside the contour, which consists of two inclined semicircle plane arches may be
prestressed
examined as an example (Fig.2).
3. C ALCULATION OF DISPLACEMENTS AND INNER FORCES OF HYPAR-NETWORK
The stress-strain State of the network is to be described by means of conditions of equilibrium and
equations of deformations compaibility |
write (Fig. 1)
1
|. For the case of action of vertical loads p we may
G *(* + »)+g<ffe + w)
(31)
dx2
dy2
GG° [i + (A)2]l.
* +
dw
dz + ±dw}
(3>2)
Etx
dx
dx dx
dx
2
dx
g"g° [1 + (^)V - * + ** <-*
+ 1*»)
(3.3)
Ety
dy
dy
dy
dy
2
dy
where w, v
- displacements in directions of axes x and y correspondingly,
G,
H
- the final horizontal forces of the network,
E - modulus of deformation of cables,
t,\
tv
- effective thicknesses of the carrying and the stretching cables correspondingly.
For elimination horizontal displacements
we may equate the network edge displacements
to
displacements of the contour beam, caused by action of horizontal forces of the network.
After approximation of the deflection function for the network and using Bubnoff-Galjorkin
procedures we get for the relative deflection (0—- a cubic equation
*x
(1
+ * +4$)Co3 +3[(1 - «*) + 2(1 - a)5]C02 + {2[(1 + a2i|r) + (1 - a)H] +
<3-4)
+ (l + a-1) (l + Xx )d + H)P0mK0 - a + xx )d + H)p*
fy
5fx2
5/v2
a4t(UXx)
where
i
a - —;
v
—JL; v
—iL.; Uj
y-
;
**
Ky
A
3a2
3*
hUx(LXj
I
+ _L
$
_ geometrical factors;
5Ety°
v
*
- parameter of contour rigidity;
F
12EcIc<l+xx)
EJt. - the bending rigidity of contour beam;
* £2Ü—
p
_ parameter of the network pretension;
lQEtxfx*
n* - ———
loading parameter
WEtxf>
V. KULBACH 41 The horizontal components of the network inner forces may be presented as
V. KULBACH
41
The horizontal components of the network inner forces may be presented as follows
G-G0 +
(3.5)
9a2(l + X*)(l+H)
5£ry//C0[(2a-C0)-|(l-a + Co)51
H-H0-
9b\\ + Xy)(\+H)
(3.6)
4. ESTIMATION OF BEHAVIOUR OF HYPAR-NETWORK
Hypar-network with elliptical contour, supported by the vertical columns, is characterizied by some
specific qualities. The contour without any outer horizontal supports may freely widen or narrow
under action of cable forces, caused by loading the network. Therefore the contour stiffness and
other system parameters has decisive import for the stress-strain State of the network. In the
following will be described some aspects of behaviour of the network and its collaboration with the
contour beam:
I) Dependence of relative deflection of the network on load parameter has obviously nonlinear
character (Fig.3), similar to analogous relationship for the plane suspended Systems;
3.
0.4
**£&&
UO
0.3
0.2
0.1
P
0.1
0,2
0,3
0,4
0,5
0,6
Fig.3 Dependence of displacements on the load parameter
2) Inner forces of carrying cables under action of network loads will increase by all values of
bending rigidity of the contour. Inner forces of stretching cables will decrease in the case of great
and increase by small bending rigidity of the contour beam. Therefore the bending moments of
flexible contour beam are comparatively small.
3) Dependence of network deflection on the parameter of contour rigidity is shown on the Fig.4.
It is remarkable, that by very great and very small rigidity of the contour beam, deflection of the
network does not change remarkably.
ML 42 DESIGN AND ERECTION OF LONG-SPAN G-G.1H-H. 700 SP $> *w 600 5 *
ML
42
DESIGN AND ERECTION OF LONG-SPAN
G-G.1H-H.
700
SP
$>
*w
600
5
*
*
»*
500
KP
u.
**
?J
ix
400
V-^
s
Ä
S-
300
/
*a2L*
200
^
x'
100
,'
-p
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
Fig.4 Dependence of cable forces on the load parameter
5.
ERECTION EXPERIENCE OF SADDLE-FORMED SUSPENDED ROOFS IN ESTONIA
Two saddle-formed network structures have been erected in Estonia as acoustic screens for song
festival
tribunes.
The first of them
was
erected
in Tallinn
in
1960.
It is inclined network,
prestressed inside the contour, formed by two plane arches and supported by massive counterforts.
The bearing structure of the acoustic screen of the tribune in Tartu erected in 1993 is an hypar-
tormed network within a contour, constructed as a spatial tubulär rod with axis, having elliptic and
parabolic projections. The contour is supported by three plane supports, connected with the contour
and foundation by linear hinges. The supports do not resist symmetrical horizontal displacements
of the contour. Therefore the interaction between the network and the contour is of particular
importance for balancing outer loads. Because of acoustic requirements the cladding of both screens
in Tallinn and in Tartu was made from timber panels. For tribune in Tartu the collaboration
between the network and timber shell is taken into account.
REFERENCES
1. KULBACH V., Hanging Roofs as Acoustic Screens for Song Festival Tribunes in Estonia.
Transactions of Tallinn Technical University, No 721 /1990.
2.
KULBACH V., Behaviour of Tubulär Contour of Hypar-formed Hanging Roofs. Tubulär
Structures. London and New York, Elsevier Aplied Science, 1990, p.437-432.
The Lightest Retr Toiture escamotab Das leichteste einz Matthys LEW Principal Weidlinger Associates New York,
The Lightest Retr
Toiture escamotab
Das leichteste einz
Matthys LEW
Principal
Weidlinger
Associates
New
York, NY, USA
PWl «
t
SUMMARY
Tenstar Domes with cable networks rigidized
to a wide variety of configurations. For the G
plan was covered with a teflon coated fiberglas
Tenstar Dome. Alternate configurations have
much as 360 m and one scheme includes a re
usefulness of a facility by providing all-season
44 LIGHTEST RETRACTABLE ROOF 1 Introduction. The first historical retractable roof was the canopy over
44
LIGHTEST RETRACTABLE ROOF
1 Introduction.
The first historical retractable roof was the canopy over the coliseum in Rome and was more like a
horizontal curtain hung from a series of parallel ropes strung between the top of the stage house
and the back of the Stands. It served more as a sun shade rather than a weatherproof cover because
of its numerous open joints. Thirty years ago, a modern retractable roof was built in Pittsburgh
consisting of a number of orange peel sectors supported along the edge on a circular track and in
the center of the arena, from a cantilevered arm. The roof, apart from being very expensive,
suffered from mechanical problems and is now permanently shut. When the Toronto Sky dorne was
completed in 1988, modern technology and sophisticated Computer analysis dealt successfully with
the mechanical difficulties that had plagued the Pittsburgh structure and its temperature and
deformation dependent deformations. However, the cost issue remained as that structure exceeded
$400 million dollars. The Toronto Sky dorne and recent Japanese retractable domes have been
constructed as rigid, heavy, steel structures, either spanning across an arena or cantilevering from
the edge.
The Tenstar retractable dorne1 offers another approach based on the principle that a cable structure
that remains in place over the opening is virtually invisible and totally transparent. By introducing a
structure that supports a retractable roof, a lightness of structure is achieved that is directly
proportional to a "lightness" of cost.
1.2 The System.
The first Tenstar Dome was completed in Atlanta in 1992 and consists of a 240m by 193m fabric
covered oval dorne. It was based on the tensegrity concept proposed in 1954 by the American
original, Buckminster Füller. This structure consisted of ever smaller annular rings, rigid in their
Patent Penchng
#4 M. LEW 45 vertical planes, connected to each other with cables running from the
#4
M. LEW
45
vertical planes, connected to each other with cables running from the top of the large ring to the
bottom of the next smaller ring. He described it as a structure in which islands of compression
reside in a sea of tension. Structurally, it can be described as a radially oriented succession of
discontinuous trusses in which the bottom chord is a series of hoops tying together all the trusses.
In the Tenstar Dome, each node is braced by triangulated cables forming, on the top surface, a
continuous net. The resulting arrangement is an extremely stiff structure in which the stiffness is
obtained both from the triangulation and the prestress necessary for a cable net. The total dead
weight of the resulting structure, including its fabric cover is an incredibly low 0.3 kN/m2, less than
one half the wind suction load specified by most building codes as well as the required live load.
SWS
The system allows and incredible variety of alternative configurations: the plan may be a circle, a
curved triangle, or an oval; the covering material may be fabric or a rigid metal deck; openings in
the roof may be introduced leaving only an annular ring to cover grandstands and most exciting of
all, a retractable roof can be devised.
2.1 The Retractable Option.
The retractable Option we have studied offers the lightest weight and concurrent lowest cost roof of
this type developed to date. The scheme consists essentially of a cable dorne with parallel cables on
the top surface to which parallel tracks are attached. The spacing of the cables is of the order of
15m which renders this sparsely populated cable grid relatively invisible to the spectator looking up
at the sky.
Nevertheless, the spacing is small enough to permit a lightweight truss structure to be
designed to ride on it. This movable roof is assembled from a series of separated triangulär trusses,
linked to each other in the direction of travel in the manner of a Caterpillar. In the orthogonal
direction, the sections are connected with spring loaded rods to allow rotation and some lateral
displacement. This concept permits the moving roof to adapt to a deformable cable supporting
structure.
ML 46 LIGHTEST RETRACTABLE ROOF By accepting the minimal Visual impingement of a cable net,
ML
46
LIGHTEST RETRACTABLE ROOF
By accepting the minimal Visual impingement of a cable net, an extremely lightweight movable roof
structure is possible compared to the heavy, trussed, and often cantilevered movable roofs that have
been built.
2.2
Alternative Configurations.
Two types of these retractable roofs have been studied: An oval roof with a rectangular retractable
oculus; a round roof with four petal-like retractable sectors. In both cases, the retractable sections
are arranged for the greatest simplicity and ease of Operation and maintenance.
Because the retractable roof sections are lightweight, the bogies and operating mechanisms
including wheels, housings, motors are all scaled accordingly and are lightweight. Two operating
Systems have been considered: A cable driven system with fixed motors located on the compression
ring. A wheel driven system with built-in small horsepower electric motors directly geared to the
wheels on the operable section.
At the meeting Stiles of the roof sections, a weatherproof interlocking device has been proposed.
As the sections approach each other, a cam displaces a spring loaded cap on one section that then
overlaps a curb on the other section.
WHEEL
HOUSING
WHEEL
SIDE
FLASHING
r
AXLE
UPLIFT RESTRAINT
BLOCK
WITH
TEFLON
COATED
MEETING
"^ f h
SURFACE
CABLE
CLAMP
ISOMETRIC
VIEW
WHEEL HOUSING
DETAIL
M. LEW 47 The fixed portion of the roof can be covered in fabric using
M. LEW
47
The fixed portion of the roof can be covered in fabric using the hyperbolic paraboloid panels used
on the Georgia Dome or can have a rigid covering of metal panels supported by steel joists. The
movable roof panel can similarly be covered in fabric using saddle shaped panels between Sprung
arches or can have a rigid metal covering. There is an inherent advantage in using fabric as the roof
covering because it obviates the need for a separate waterproofing membrane that, because of the
flexible nature of the roof, needs flexible joints, all of which are sources of potential leaks. On the
other hand, in locations where insulation and the need to support large snow loads are important, a
rigid roofing material may be more appropriate.
ROOF CABLE LAYOUT PLAN
OPENED POSITION
CROSS 8ECTIOW
Lee Bla

Lee

Bla

Ban Seng CHOO Civil Engineer Univ. of Nottingham Nottingham, UK
Ban Seng CHOO
Civil Engineer
Univ. of Nottingham
Nottingham, UK

Retractable Roof Using th

"Cadres reciproques" pour le

"Wechselseitige Rahmen" f

Paula N. COU

Civil Engineer

Univ. of Nottin

Nottingham, U

50 RETRACTABLE ROOF USING Jlk L LNTRODUCTION Across the globe architects and engineers are exploiting
50
RETRACTABLE ROOF USING
Jlk
L LNTRODUCTION
Across the globe architects and
engineers are exploiting the
versatility of retractable roofs for
sports stadia and arenas. In North
America, the Toronto Skydome
[1,2],
the
Olympic Stadium
in
Montreal [2], and the Civic Arena in
Pittsburgh [2], are well known examples
of
structures
with retractable roofs.
The
structural design
of
each one is
unique.
In the
Skydome,
three
separate roof sections retract and nest
over the top of a fixed panel; each panel is a different size
and shape from the others. The Montreal
Olympic
Stadium has a fabric roof that can be roiied
up
into atower but because of difficulties
in
retraction, it now remains
permanently
retracted.
The Civic Arena in Pittsburgh
features a successful retractable dorne where six
separate
sections pivot
about a
pin
and roll along curved rails before Coming to rest over two
cantilevered
fixed sections.
Additionally, there are several structures which
Mukogawa High School Swimming Pool, the Ariake
are in Operation in Japan, such as the
Colosseum, and the Fukuoka Dome [3].
Apart from
these buildings, which have successfully opening
and closing roof Systems, there
are
other
many
designs being developed
for
retractable roofs on
sports complexes all over the
world.
One such
design
being
worked on at the
University of
Nottingham is the Reciprocal
Frame Roof [4], see Figure 1.
A related reticulated
roof
in
the form
of
a spherical dorne was
proposed by
Emilio Perez Pifiero in his
patent
9102733 of 1961
[5].
The likeness of these two
structures
to the iris of a camera shutter makes them possible Solutions for retractable roof
Systems.
2. THE RECIPROCAL FRAME ROOF
2.1 Architectural Aspects
As the Reciprocal Frame Roof has the
same conflguration as a camera shutter, it is
easy
to
visualize its opening and closing.
The static
conflguration has considerable Visual impact and
appears
very
dynamic when viewed from the floor.
The
primary beam structure seems to be
rotating
about an axis, in
empty
at the centre of the
roof.
space,
During the retraction
process,
like
the
leaves of an iris diaphragm, each beam rotates
individually about its external support,
opening and closing
the structure, see
Figure 2 A to D.
A
major
advantage to using
the
Reciprocal Frame
Roof as a retractable roof
system is the versatility
in
floor
plan
design
as this
is not limited by the need to provide extensive
running rails (either in a straight line or constant
curve) that are required by most other Systems.
structure need not have the same shape, this
Since the outer and inner
polygons in the
roof can cover Stadiums with
virtually any
geometry.
2.2 Structural Components of the Roof
The primary structure of the Reciprocal Frame is a circuit of beams
spiralling
around an
imaginary centre. Each beam in
the grillage both supports and in turn is supported by the other
beams in the structure, hence the
name reciprocal.
The beams are placed tangentially around a
central closed curve so that they rest
upon the preceding beams creating
the closed circuit. An
enclosed polygon
is formed with a set
of
radiating beams equal to the
number of sides of the
polygon.
The
outer ends of the beams in the
grillage
are
supported by columns or walls.
These columns or support positioning on the walls also form a
polygon
but its shape need not
be the same as that
of
the outer to the inner
the inner polygon. As the beams rest on each other, there is a rise from
polygon which depends on the height between beam centre lines at their
intersection, number of
beams, etc., creating a three dimensional structure.
2.3 Structural Behaviour
The conflguration of the beams in the Reciprocal Frame
Roof causes applied vertical
loading
to
be converted
into downward thrust acting through the outer
supports.
In a regulär circular or
polygonal
form
carrying
a uniformly
distributed vertical load, the beam reactions are all
equal
to the
total roof load divided
by the
number of beams. However, when a point load is applied
B.S. CHOO, P.N. COULIETTE, J.C. CHILTON 51 Fig.l Three-dimensional view of Reciprocal Frame Fig.2 Plan
B.S. CHOO, P.N. COULIETTE, J.C. CHILTON
51
Fig.l Three-dimensional view of Reciprocal Frame
Fig.2 Plan views of retractable Reciprocal Frame structure
to an individual beam it will be
partially
carried
by
all the beams in the
grillage and individual
beam reactions will depend on the
position of the load and
geometry
of the
grillage. Any load
applied to the structure affects all beams and similarly the
deflection of
one beam produces
displacement in all beams.
Since the elements
of
the Reciprocal Frame are resisting
the loads
by bending, no central compression
ring
or external
tension ring is required.
When the roof
retracts,
it
is
simply
changing the
shape of the inner
polygon
and therefore similar load
distributions and
characteristics occur.
These features are
particularly
interesting when the
structure is
subjected to wind and snow loading. Non-symmetrical loads,
for
instance, a snow
load on one side
of
the roof, are distributed more evenly throughout the structure.
Jt 52 RETRACTABLE ROOF USING 3. ROOF OPERATION 3.1 Geometry of Opening/Closing System To open
Jt
52
RETRACTABLE ROOF USING
3. ROOF OPERATION
3.1 Geometry of Opening/Closing System
To
open
the Reciprocal Frame Roof, each beam rotates
about its outer support.
The degree to
which
each rotates depend
the number of beams in the
upon
grillage
and the
shape of the inner
and outer polygons. However,
if the beams are of the same lengths,
it is important that they all
rotate simultaneously and by
the same amount about their outer
supports.
Due to the
complex
three-dimensional
geometry,
displacement of one beam greatly affects the shape and structural
integrity of the roof.
As
the roof
opens, each beam rotates about its outer
support,
both
vertically
in the
plane
of the beam and
horizontally
toward the outer polygon. Since
the beams
are all
interdependent, it is important that they are
always
being supported by the previous
beam. Opening
the roof moves the inner
support
of the beams
along
a curved path towards the
outer
support,
as shown in Figure 3, therefore, the beam must be at least as long as the
maximum
distance to the support.
This distance is
only
slightly
longer than each beam but the
exact distance depends on the
shape of the polygons and number
of
beams in the structure.
3.2 Drive Mechanism
The beams of the roof can rotate independently using individual synchronized
motors or can be
connected by an outer ring which is mechanically rotated, in turn rotating
each of the beams
through the
required
distance.
However, use
of
such a
ring
greatly
inhibits the amount of
opening
in the
roof
because the radius of rotation of the beam and the
radius of the entire roof
(radius
of
ring)
are different. To account for this difference, it is possible to insert a connecting
link between the
ring
and the beam at an
angle
which turns the beam without
requiring
a large
rotation of the ring.
Without the ring, the need for massive track structures
(as in
the
Skydome, etc.) disappears.
3.3 Joints and Supports
The beams are connected to the outer
supports
using a hinge
which allows for vertical and
horizontal rotations.
If tapered beams are
used, the vertical
rotation of the beam could be
greatly reduced or possibly
eliminated.
The inner
joints
must allow for both continual
support
of
the preceding beam
and the movement of both beams. To accommodate this motion and
maintain support,
a rolling
Joint must be used.
It is important to note that the Joint does not
connect to
either beam but
simply
rolls between the
two,
acting
only as
a guide for the
retracting movement and maintaining the position of support required.
4. DESIGN
4.1 Design Advantages
There are several
advantages
in
using
the
Reciprocal
Frame Roof as a retractable roof structure.
First of all, it or equivalent structures, such as a
salt storage building of 26 metre span made of
11 tapered, glued
laminated beams using an analogous planar grillage as reported
by Natterer,
have already
been successfully used as a static roof structure [6].
Also, because
of
its unique
load distribution there is no need for internal
support
and, therefore, can be used in large arenas
and stadia. Another benefit to the structure is that
the strength is in the principal design and not
a specific material allowing freedom in choosing the building material. The 'beams' discussed
in this article refer to the basic structural members. These
members can actually be steel,
timber or concrete beams or steel or timber trusses, or even planes of
frame. Another
space
benefit to this design is that it is not limited by shape. Both the outer
polygon and inner
polygon can be
of
any shape to suit the need. Because they
need not have the same shape, the
roof
can
be
used
to cover oddly shaped buildings
and still maintain its retractable
characteristics.
#Il B.S. CHOO, RN. COULIETTE, J.C. CHILTON 53 Fig.3 Curved path of intersection point of
#Il
B.S. CHOO, RN. COULIETTE, J.C. CHILTON
53
Fig.3 Curved path of intersection point of beams
4.2 Details of Design
Because the retractable
Reciprocal
Frame is a new
concept, which has not
yet
been considered
for use in a "real" project, specific details have not
been addressed completely.
However,
these details have not been completely overlooked.
Such
topics as roofing material, drainage,
and other potential difficulties in the design have been reviewed.
The Reciprocal Frame Roof allows for
flexibility
in choosing the cladding material.
If static
panels are used, the beam and panel would
act as
a plane in retraction.
It
is also possible to use
a fabric membrane which folds when the roof is
and Stretches out when the roof is shut.
Another
design
consideration in
making
this
open
roof structure practical is drainage.
However, the
roof has a
natural slope to it and the Solution
would be found in directing the
flow of water to a
Strategie point.
As with
any
roof, the use of
proper
detailing
and sealants would eliminate
leaking.
The most obvious
of the obstacles to overcome in the Reciprocal
Frame Roof is the
hole in the centre which is formed because the beams can only completely
close theoretically if
they
have no thickness. This
problem however, can easily be solved by attaching small panels
to
the sides of the beams, at the
central polygon, which come completely together when the
beams are in their closed position.
A further consideration is that of disproportionate collapse. As each beam depends on all of the
others for its
support,
the removal
of
or
damage
to
one beam
any
may
result in collapse of the
whole structure
unless it is appropriately designed.
One
of
way
overcoming this potential
weakness is to provide suitable
supports
and to design the beams to act as cantilevers under the
reduced
loading required
for the
accidental limit State whilst
utilising
the reciprocal
action for
füll ultimate
loading.
However, if the opening leaves of the
roof
are construeted
from
Space
frames, the structural stability does not rely totally on individual beams and the problem
of
disproportionate collapse can be avoided.
54 RETRACTABLE ROOF USING Jrm 5. CONCLUSIONS In conclusion, the Reciprocal Frame Roof has considerable
54
RETRACTABLE ROOF USING
Jrm
5. CONCLUSIONS
In conclusion, the Reciprocal Frame Roof
has considerable potential for use as a retractable
covering for Sports arena and stadia. It has a
powerful geometry
with a
dynamic Visual impact.
The design
is
with
very
versatile and can be used
a wide
variety of
floor plans.
Some
examples
of which
can be seen in Figure 4.
Fig.4 Reciprocal Frames over various plan shapes
REFERENCES:
ROORDA, J.,
SHERBOURNE, A.N., SRIVASTAVA, N.K.,
"Toronto Skydome"
Proceedings
from the International Association for Shell and Spatial S tr u et u es
Canadian
Society of Civil Engineers International Congress on Innovative
Large Span
Structures, Toronto, Canada, July 1992, Volume 1, pp.51-115.
O'CONNER, LEE., "Toronto Skydome: The House that Engineers Built." Mechanical
Engineering. October 1992, pp.54-57.
NARITA, H., "Examples
of Retractable Roofed Domes in Japan" Proceedings
of the
International
Association for Shell and Spatial Structures - Canadian Society
of Civil
Engineers
International Congress on
Innovative Large Span Structures, Toronto,
Canada,
July 1992, Volume 1, pp.498-509.
4. CHILTON, J.C. and CHOO B.S., "Reciprocal Frame Long Span Structures" Proceedings
of
the
International Association for Shell and Spatial Structures - Canadian Society
of
Civil Engineers International Congress on Innovative Large Span Structures, Toronto,
Canada, July, 1992, Volume 2, pp. 100-109.
5. ESCRIG, F.
"Las Estructuras de Emilio Perez Pinero" in Arquitectura Transformable,
Textos de Arquitectura, Escuela Technica Superior de Arquitectura de Sevilla, Spain
1993, pp. 30-32.
6. NATTERER, J., HERZOG, T. and VOLZ, M.
Holzbau Atlas Zwei. Institut für
Internationale Architektur, Munich, p. 179.
55 Developments in Cold Formed Roof Structures Developpements dans les toitures structurales ecrouies ä froid
55
Developments in Cold Formed Roof Structures
Developpements dans les toitures structurales ecrouies ä froid
Entwicklungen in kaltverformten Dachtragwerken
John C. CHAPMAN
Director
Robert PRINGLE
Product Manager
Patrick DOWLING
Prof.
Chapman & Dowling Assoc.
Haywards Heath, UK
Space Decks Ltd
Chard, UK
Imperial College
London, UK
John
Chapman
was
Robert
Pringle spent
Patrick
Dowling
is
successlvely Reader in
Structural Engineering
fifteen
years
in
site
Head
of
Civil
engineering
and project
Engineering at Imperial
at
Imperial
College,
management
for
a
College
and
is Vice-
Director
of
the
variety
of
major civil
Chancellor and Chief
Constructional
Steel
engineering
and
Executive
of
Surrey
Res.
and
Dev.
building
projects,
in
Univ.
He is a Director
Organisation,
and
four
continents. For the
of Chapman & Dowling
Group
Techn.
Director
last ten
years
he has
of George Wimpey plc.
been
engaged
in
Assoc. He has been
Chairman of the lead
He
is
a
Visiting
product
development
committees
dealing
Professor at Imperial
and
construction
with
EC3
since
its
College, and a Fellow
management
of cold
inception.
He
is
a
of the Royal Acad. of
formed
structures.
Fellow
of
the
Royal
Engineering.
Acad. of Engineering.
SUMMARY
The advantages of cold formed sections are discussed, with
particular reference to two
space
frame roof Systems. Designs methods which are used for lip
buckling,
flexural
torsional, and lateral torsional buckling are outlined. The transfer of forces and moments
at the connections
is described. The ränge of structures in which the system have been
applied is indicated.
RESUME
La
communication presente les avantages des sections profilees et formees ä froid, en se
referant ä deux
systemes
de toiture tridimensionnels. Les auteurs
exposent
les grandes
lignes des methodes
de
calculs utilisees
pour
determiner le voilement
de bordure, la
torsion
par
flexion et le flambement
par
torsion en flexion. Ils examinent en outre le
transfert
des forces et des moments sur
les assemblages. Ils presentent la gamme de
structures pour lesquelles ces systemes ont ete appliques jusqu'ici.
ZUSAMMENFASSUNG
Diskutiert werden die Vorteile kaltverformter Profile im
Zusammenhang mit zwei
Raumfachwerk-Dachsystemen. Dabei sind die Bemessung für
Randbeulen, Biegetorsion
und
Biegedrillknicken und die Einleitung von Kräften und Momenten an den Anschlüssen
wichtige Fragen. Die Breite der bisherigen Anwendung auf Tragwerke wird umrissen.
56 DEVELOPMENTS IN COLD FORMED ROOF STRUCTURES 1. INTRODUCTION As techniques of structural analysis, powerfully
56 DEVELOPMENTS IN COLD FORMED ROOF STRUCTURES
1. INTRODUCTION
As techniques of structural analysis, powerfully assisted by rapidly developing Computer hardware
and Software, reach maturity, the scope and economy of construction depends increasingly on
developments in manufacturing methods, construction equipment, and materials.
with its associated processes, is opening new avenues for structural design.
Cold forming,
Adjustable cold rolling mills now enable purpose designed sections to be produced from a
Stockholding of steel Strip.
Holes are pre-punched on the rolling line, members are cut to length
by flying shears, and when required, powder coating is then applied. Dimensional tolerances of ±
lmm, and bolt diameter clearances of 1mm are normal. Sections with a perimeter length of up to
1.2m and up to 8mm thick can now be rolled.
Pre-galvanised strip with a yield stress of 390
N/mm2 is available up to 5mm thick/ The cold rolled components for a roof structure covering
5000 sq m can typically be produced in 5 working shifts. Components can usually be man-
handled. Costs per unit weight are somewhat greater than for hot rolled sections, but this is offset
by weight reduction, corrosion protection and response time.
In addition to the production advantages, the designer has the freedom to exercise ingenuity in
devising Optimum sections to meet architectural and structural requirements.
The structural
analysis of cold formed members is however considerably more complex than for hot rolled
members, because of the various forms of instability which must be considered. Many more
developments in production methods, in design, and in applications, can be expected.
The investigations which provide the background for this paper have centred around two patented
Systems, but the approaches and methods which have been developed have general application.
2. THE HARLEY SYSTEM 80
The Harley Joint for a square on square space frame is shown in Figure 1. The guiding principle
of the Harley system is to minimise the cost of the nodal connection.
Chord splices are located
away from nodes, and occur every three or four modules. Joint economy is achieved at the cost of
introducing end moments to the chords.
The moments are much greater in the inner chords, and
are a maximum in the vicinity of columns. The provision of moment capacity across nodes does
however exclude the possibility of toggle failure, and enhances the robustness of the structure.
The number of components is minimised by bending the flattened ends of the tubulär diagonals to
enter between the back to back Channels, so forming an architecturally neat connection.
The
creases of compression tubes are not fully supported;
this limits the diameter of the tube which
can be used, and causes an end eccentricity. Very formable steel is required, and during assembly
the tubes are supported only at the lower ends. The length of the end tab is limited by the width of
the section, so the end distance limits the tensile strength of the tube. Nevertheless, the system has
proved to be economic, and many structures have been built using the Harley system [1].
3. THE MULTIFRAME SYSTEM
This system also uses back to back Channels, but the diagonals are connected through wing plates,
and the flattened ends of the tubes are not bent (Figure 2).
The tab length is not limited by the
width of Channel, more than one bolt can be used if required, the end distance is not limited, and
J.C. CHAPMAN, R. PRINGLE, P. DOWLING 57 the end eccentricity of the diagonals is very
J.C. CHAPMAN, R. PRINGLE, P. DOWLING
57
the end eccentricity of the diagonals is very small. The wing plate transmits the resultant force of
the four diagonals to the nodal connection, and leaves no voids between the back to back Channels.
The greater. eccentricity moment is in the longitudinal direction of the wing plate, which therefore
assists the transfer of this moment, by virtue of the bending strength of the trough. The cost of the
additional components is offset by increased structural efficiency, and by removing the demand on
formability imposed by bending the tabs. Also the tube size is not limited, and much greater
spans
are possible than with the Harley system. The system can also be applied to planar, triangulär, and
box trusses, and to portal frames.
TOP NCDE
TOP NCDE
Node Bolts
Washer Plate
Node Bolts
Ö>
Upper Top Chord
Washer Plate
Wing Plate
Upper Top Chord
Lower Top Chord
Tube Diagonal
Tube Diagonal
Lower Top Chord
Washer Plate
Washer
Plate
\
— Upper Bottom Chord
TCP NODE
Upper Bottom Chord
^
——'Tube Diagonal
Tube Diagonal
Wing Plate
Lower Bottom Chord
Washer Plate
Lower Bottom Chord
Node Bolts
Washer Plate
Node Bolts
.BOITOM NODE
BOITOM NODE
Fig.l. Harley connection
Fig.2. Multiframe connection
4.
STRUCTURAL BEHAVIOUR
The following paragraphs outline the phenomena which must be considered in design.
A general
description of the development is given in [2] and details can be found in [3].
4.1
Local buckling
The forces and moments are computed assuming fully effective sections.
Section strength checks
are then made taking account of loss of effectiveness due to local buckling.
The effect of stress
level in an element on the effective width may be taken into account. The effect of coupling of the
sides of the section may also be considered in calculating the critical stress.
4.2 Lxp buckling
The lips of the Channel can buckle laterally as struts on an elastic foundation, the "foundation"
stiffness being provided by the transverse deformational stiffness of the section (Figure 3).
An
important question
to consider
was
whether local
buckling
of thin
walled
sections
would
58 DEVELOPMENTS IN COLD FORMED ROOF STRUCTURES significantly reduce the deformational stiffness of the section.
58 DEVELOPMENTS IN COLD FORMED ROOF STRUCTURES
significantly reduce the deformational stiffness of the section. Finite element analysis confirmed
that for typical
lip/local buckling wavelength ratios (about 7:1), the effect of local buckling on lip
insignificant.
buckling resistance can be found
by assuming
an initial
buckling is
The lip
deformation (related to manufacturing tolerances) which is affine to the critical lip buckling mode,
and using a modified Perry equation [4].
The effect of lip buckling on flexural or torsional
flexural buckling can then be taken into account approximately by reducing the yield stress to the
lip buckling resistance.
There is some evidence however that when torsional buckling develops
before lip buckling, so that both ups buckle laterally in the same direction, in a Single half wave,
lip buckling is thereby inhibited.
There are two reasons for this - the lip buckling stress for a
Single half wave is much greater than for the preferred mode, which typically has three or four half
waves, and when the lips buckle in the same direction, the foundation modulus is increased.
Systematic studies are now in progress to study, quantify and establish criteria for this interactive
behaviour.
^-
Fig.3. Development of Lip Buckling
4.3 Torsional buckling
In current design codes, torsional or torsional flexural buckling is treated by substituting the
critical torsional buckling stress for the Euler buckling stress in the Perry equation for imperfect
struts.
In fact the torsional buckling stress distribution (Figure 4) is quite different from that for
flexural buckling, and can be found by applying Young's equation for the deflection of imperfect
struts separately to the translation and rotation of a member buckling torsionally[5].
It was also
shown that a simple modification to the Perry equation, which then correctly represents torsional
buckling of a doubly Symmetrie section, provides a satisfactory approximation for torsional
flexural buckling. The section shown in Figure 3 is prone to lateral torsional buckling, which also
can be treated by the method just described.
J.C. CHAPMAN, R. PRINGLE, P. DOWLING 59 Initial perfect Fig 10 ^Imperfect FINASIC 4= ¦h
J.C. CHAPMAN, R. PRINGLE, P. DOWLING
59
Initial perfect
Fig
10
^Imperfect
FINASIC
4=
¦h
JgT^
CG
*f
CG
"*+-.
i—
Lw_
/
f^^ä Compression
I
«
Compression
rT^
-SC*
I
I Tension
I
I Tension
/
i'sc
Fig.4. Comparison of Torsional buckling stresses - design model and finite element analysis
Torsional flexural and lateral torsional buckling interact strongly, because both buckling modes
consist of rotation and translation of the section, the member buckling in a Single half wave.
When moment causes lip tension, torsional flexural buckling is inhibited to an extent which
depends on the magnitude of the moment It may conservatively be assumed that the resistance is
equal to the torsional flexural buckling resistance, or to the flexural buckling/moment interaction
value, whichever is smaller.
Similar considerations arise in respect of interaction between axial
tension and moment which causes lip compression. The interaction between the buckling modes
is an important theme of current research.
4.4 Nodalforces
In the Harley system, vertical forces are transferred at the tube creases to the washer plates, which
are held together by four bolts, which resisf tension and shear.
The washer plates transfer
components of the horizontal tube forces as axial forces applied to each chord. The washer plates
also transfer components of the eccentricity moments to the chord tables and thence to the chord
webs, which are therefore subject to vertical forces which vary from tension to compression over
the length of the washer plates.
The table of the chord is subject to varying vertical shearing
forces between the edges of the washer plates and the chord webs. Where necessary, nodal inserts
are provided, which prevent web crippling and table shearing. The crease of a compression tube is
not supported over its whole length, and the tube
size must be limited
to prevent crease
deformation, and also to limit the eccentricity moment applied to the tube. The crease of a tension
tube is fully supported, and the strength is determined by the tab bearing stress, which is limited
by the end distance from the bolt hole.
In the Multiframe system, the wing plates and washer plates transfer the resultant forces and
moments to the chords.
The greater moment is transferred over a length measured from one end
of the wing plate to the other end of a washer plate, so the vertical Channel web forces are reduced.
The flattened ends of the tubes remain straight, which simplifies fabrication and places much less
demand on formability.
The tubes are axially loaded, a more efficient end distance can be
one bolt can be used if required,
The wing plates must have sufficient
adopted, and more than
resistance to bending about the mid-thickness. Bending is caused by eccentricity, and by
misalignment, which results from manufacturing tolerances and from Joint translations and
rotations due to load.
60 DEVELOPMENTS IN COLD FORMED ROOF STRUCTURES In any system employing flattened tubes, if the
60 DEVELOPMENTS IN COLD FORMED ROOF STRUCTURES
In any system employing flattened tubes, if the strength is not limited by the tensile connection,
flattening
of the tube can occur before the tensile strength of the tube is attained. In a compression
tube
which is short enough not to buckle, outward bulging of the end regions can occur, if the tube
is sufficiently thin.
5. DESIGN, FABRICATION AND CONSTRUCTION
It will be apparent from the above discussion that many design checks must be made, in addition
to the frame analysis. Remembering that many more structures are designed in outline than are
actually built, the necessity for computerising the design process is apparent. In fact the
development of these Systems depends as much on the Computer as on the manufacturing process,
which itself is Computer controlled. The Computer executes and couples analysis, design,
estimating, scheduling, and the control of manufacture.
The high rate of production necessitates careful pre-checking of component schedules and
dimensions.
The dimensional accuracy which is innerem in the production system ensures rapid
assembly on site.
Roof structures are normally assembled at ground level. In the Multiframe system, the bottom
chords are laid out on level stools, and the wing plates are inserted and bolted. The diagonals and
top wing plates then provide statte pyramids on which the top chords can be laid and bolted.
Large roof structures are raised by special lifting columns, which ensure that all lifting points are
maintained at the same level. This system also has advantages on restricted sites.
6. APPLICATIONS
The Harley system can provide spans with perimeter columns up to about 40m, and internal spans
between Single columns up to about 26m.
The Multiframe system does not have specific limits,
but is economically viable for perimeter supported spans up to at least 70m and internal spans up
to at least 40m. The Systems have been applied to a wide variety of roof structures and canopies,
including sports halls, transportation terminals, industrial buildings, and retail facilities.
7. REFERENCES
(1)
CODD B
E, WHITE
S
H.
Eagle Centre Market refurbishment, Derby.
The Structural
Engineer. Volume 70, No.5,1992.
(2) CHAPMAN J C, BUHAGIAR D, DOWLING P J. Developments in cold formed space
structures. Conference on space structures, Guildford, Surrey, September 1993.
(3) BUHAGIAR D. Behaviour and design of structures using thin-walled cold-formed sections.
PhD thesis, Imperial College, London 1993.
(4) BUHAGIAR D, CHAPMAN J C, DOWLING P J. Design of C-sections against deformational
lip buckling. Eleventh international Conference on cold-formed structures. University of
Missouri-Rolla, October 1992.
(5) CHAPMAN J C, BUHAGIAR D. The application of Young's buckling equation to design
against torsional buckling. Proc ICE August 1993.
Composite Lon Poutres mixtes d Weitgespannte V W. Samuel EASTERLING Assoc. Prof. Virginia Tech Blacksburg,
Composite Lon
Poutres mixtes d
Weitgespannte V
W. Samuel EASTERLING
Assoc. Prof.
Virginia Tech
Blacksburg, VA, USA
\ ^ ^
W. Samuel
Easterling,
born 1959, graduated at
West Virginia Univ. and
V
Iowa State Univ. He has
1
been
involved
in
re-
i*
search
of
composite
floor
Systems
for
the
past 10
Now at
years.
Virginia Polytechnic Inst,
and State Univ. his re-
search
in
composite
floor system is ongoing.
62 COMPOSITE LONG-SPAN JOISTS 1. INTRODUCTION Such a The benefits bf open, essentially column free
62
COMPOSITE LONG-SPAN JOISTS
1.
INTRODUCTION
Such a
The benefits bf open, essentially column free floor Spaces in buildings are well recognized.
conflguration provides maximum flexibility in leasable space arrangements and thus gives an owner
the ability to easily accommodate the requests
of new tenants.
Composite long-span joists (light
trusses) are one structural system that provides
these large open areas.
In addition to the benefits of
column free space, the open
web conflguration of the joists permits easy access for mechanical and
necessarily increasing the floor-to-floor heights in the building.
Both of
Service Systems, without
these characteristics are significant benefits and make composite joists an economical structural
system, as was realized in three recently constructed projects in the United States.
The use of composite
joist
Systems in the U.S. is hampered by the lack of a design specification. At
present,
if a structural engineer wishes to consider a composite joist alternate, there are two
scenarios available.
One is for the engineer to use the available literature and design specifications,
The
including those of other countries, to design a system using available cross sectional shapes.
other Option is for the engineer to request design assistance from a joist manufacturer that has some
experience in composite joist design.
Due to the optimization inherent in joist manufacturing, the
second Option will likely be the most economical.
Several buildings have been constructed using composite joist or truss floor Systems, most notably
the Sears Tower in Chicago and the World Trade Center towers in New York. However, these
composite trusses were designed by the structural engineers on the respective Jobs and were unique
In the United States, joists or light trusses are structural elements that are manufactured in
designs.
shops that are solely dedicated to this type of fabrication.
This is in contrast to a conventional
building fabrication shop.
Joists are selected from design load tables (Performance tables) developed
and approved by the Steel Joist Institute (SJI).I1!
The joist specifications used by the SJI closely
follow the American Institute of Steel Construction (AISC) allowable stress design specification. 121
Because of the efficiency and optimization in the joist design and manufacturing process, it is unusual
for joists to be independently designed by the structural engineer and fabricated in a conventional
building fabrication shop.
It is in the context of the U. S. design and manufacturing process that
composite joists are discussed in this paper.
Results
of a comprehensive research program
that
focused
on the behavior and
strength
of
composite long-span joists are highlighted in
this paper.
Also, a brief description of three building
projects that utilized the composite joist floor system are presented.
Finally, the present Situation in
the United States, with respect to composite joist design specifications, is presented.
2. RESEARCH PROGRAM
A comprehensive composite joist research program has been in progress
at Virginia Polytechnic
Institute for the past several years. The project has included destructive
tests to evaluate strength of
both full-size joists and push-out specimens, non-destructive tests to evaluate human occupant
induced Vibration characteristics and analytical studies using non-linear finite element analysis.
2.1 Full-Size Joist Tests
A series of 11 full-size composite joist tests have been conducted at Virginia Polytechnic Institute.
The joist specimens ranged in span from 12.19 - 17.07 m and in depth from 356-915 mm.
All joists
were either a Warren or Modified Warren conflguration.
The loading conflguration consisted of
eight concentrated loads placed equidistance apart, thus approximating a uniform load arrangement.
All test specimens were loaded to failure.
W.S. EASTERLING, TM. MURRAY 63 Comparisons were made between the predicted and experimental strength and
W.S. EASTERLING, TM. MURRAY
63
Comparisons were made between the predicted and experimental strength and stiffness values.
The
predicted values were made using procedures similar to those outlined by Chien and Ritchie.ß]
Modifications to these procedures were made as outlined by Gibbings, et al. f4l and Nguyen, et al.I5!
In general the results were quite acceptable, as indicated in Table 1 (see section 6. NOTATION).
TEST
H^
Wt/Wc
IQn/TH,
Stud Position
1
0.89
0.89
1.34
unknown
2
0.97
0.89
2.23
unknown
3
0.96
0.99
1.68
unknown
4
0.76
0.81
0.98
weak
5
0.92
0.69
1.27
strong
6
1.10
1.24
1.39
strong
7
1.15
1.07
1.38
strong
8
1.17
1.01
1.40
strong
9
1.19
0.88
1.37
alternating
10
0.93
1.04
1.03
alternating
11
0.93
0.83
0.99
alternating
Table 1
Summary of Composite Joist Tests
Test No. 4 was an exception in that the experimental-to-calculated strength ratio was lower than
deemed acceptable. The behavior of this specimen was due to the placement of shear connectors in
the weak position, along with having the shear connector ratio less than unity. Detailed
presentations of the tests can be found in the project reports.!4»5!
2.2 Push-out Tests
During the course of the joist test program, problems were encountered in which the shear studs
appeared to fail at lower than expected loads. These premature failures were primarily attributed to
the influence of studs placed in the weak, or unfavorable, position. This problem led to a study of
the strong vs. weak position issue,!6! which was previously identified by several other researchers.
An additional project is underway at the time of this writing in which the behavior of shear studs
placed in metal deck profiles is being further investigated. It is believed that the results of the current
study, along with those conducted previously at Virginia Polytechnic Institute and elsewhere, will
result in changes to the stud reduction factors presently used in the AISC specifications. P'7l
2.3 Occupant Induced Vibration Studies
Because of the relatively long span of composite steel joists and corresponding large open floor areas
without permanent partitions, the possibility of annoying floor vibrations caused by human activities
exists. A number of tolerance criteria are used in North America to determine if a proposed floor
design may be annoying to future occupants. However, all of these criteria were calibrated using
floor Systems with spans in the ränge of 6.5 - 12 m.
The criterion proposed by Murrayl8! was used to evaluate the three floor Systems described in the
following section where spans are 15 - 36m. All three Systems easily satisfied the criterion. In
addition, tests of the bare floor of the Nations Bank Building (see Section 3.1 for a description) were
conducted.
It was determined that the calculation methods in the Murray criterion are applicable to
such long-span composite joist floor Systems and that the predicted human response was accurate.
That is, the floor System was found to be free of annoying vibrations due to human activities.
Also,
64 COMPOSITE LONG-SPAN JOISTS no complaints have been received concerning floor motion in any of
64
COMPOSITE LONG-SPAN JOISTS
no complaints have been received concerning floor motion in any of the three completed buildings
described in Section 3.
3. BUILDING PROJECTS
Three building projects were constructed in the last several years in the United States that utilize
composite joist floor Systems. The design process for each of the three buildings was similar in that
the structural engineer worked closely with the joist manufacturer.
Designs according to the
requirements set forth by the structural engineer of record (EOR) were carried out by the joist
manufacturers' structural engineering staff and subsequently approved by the EOR.
3.1 Nations Bank Building
The Nations Bank Building, originally the Sovran Bank Building, is located in Knoxville, Tennessee.
The 11 story building was designed by the office of Stanley D. Lindsey and Associates, Ltd. of
Atlanta, Georgia.
Composite joists are used on floors 6-11 in the office rental space, with levels 1-5
being a concrete parking structure.
The coniposite joists have a depth of 1,016 mm and are used for spans of 18.9 m.
The joist spacing
is 2,540 mm. The composite slab consists of a 76 mm deep x 0.91 mm thick composite steel deck
with a structural lightweight concrete fill, yielding a total slab thickness of 160 mm.
Joists were
fabricated using steel with a nominal yield stress of 345 MPa and have an approximate mass
per
length of 50 kg/m.
A total of 32 welded headed studs (19mm x
132 mm) were used as shear
connectors along each joist. A more complete description of the project is given by Swenssonf9!.
3.2 312 Elm Street Building
The 312 Elm Street Building is located in Cincinnati, Ohio, for which the structural design was done
by the office of Stanley D. Lindsey and Associates, Ltd. of